Editor’s note: This Medical Grand Rounds was presented as the 14th Annual Lawrence “Chris” Crain Memorial Lecture, a series that has been dedicated to discussing kidney disease, hypertension, and health care disparities in the African American community. In 1997, Dr. Crain became the first African American chief medical resident at Cleveland Clinic, and was a nephrology fellow in 1998–1999. Dr. Nally was his teacher and mentor.

African Americans have a greater burden of chronic kidney disease than whites. They are more than 3 times as likely as whites to develop end-stage renal disease, even after adjusting for age, disease stage, smoking, medications, and comorbidities. Why this is so has been the focus of much speculation and research.

This article reviews recent advances in the understanding of the progression of chronic kidney disease, with particular scrutiny of the disease in African Americans. Breakthroughs in genetics that help explain the greater disease burden in African Americans are also discussed, as well as implications for organ transplant screening.

ADVANCING UNDERSTANDING OF CHRONIC KIDNEY DISEASE

In the 1990s, dialysis rolls grew by 8% to 10% annually. Unfortunately, many patients first met with a nephrologist on the eve of their first dialysis treatment; there was not yet an adequate way to recognize the disease earlier and slow its progression. And disease definitions were not yet standardized, which led to inadequate metrics and hampered the ability to move disease management forward.

Standardizing definitions

The situation improved in 2002, when the National Kidney Foundation published clinical practice guidelines for chronic kidney disease that included disease definitions and staging.¹ Chronic kidney disease was defined as a structural or functional abnormality of the kidney lasting at least 3 months, as manifested by either of the following:

• Kidney damage, with or without decreased glomerular filtration rate (GFR), as defined by pathologic abnormalities or markers of kidney damage in the blood, urine, or on imaging tests

• 

  GFR less than 60 mL/min/1.73 m², with or without kidney damage.

A subsequent major advance was the recognition that not only GFR but also albuminuria was important for staging of chronic kidney disease (Figure 1).²

Developing large databases

Surveillance and monitoring of chronic kidney disease have generated large databases that enable researchers to detect trends in disease progression.

US Renal Data System. The US Renal Data System has collected and reported on data for more than 20 years from the National Health and Nutrition Examination Survey and the Centers for Medicare and Medicaid Services about chronic and end-stage kidney disease in the United States.

Cleveland Clinic database. Cleveland Clinic has developed a validated chronic kidney disease registry based on its electronic health record.³ The data include demographics (age, sex, ethnic group), comorbidities, medications, and complete laboratory data.⁴

Alberta Kidney Disease Network. This Canadian research consortium links large laboratory and demographic databases to facilitate defining patient populations, such as those with kidney disease and other comorbidities.

Kaiser Permanente Renal Registry. Kaiser Permanente of Northern California insures more than one-third of adults in the San Francisco Bay Area. The renal registry includes all adults whose kidney function is known. Data on age, sex, and racial or ethnic group are available from the health-plan databases.

DEATHS FROM KIDNEY DISEASE

The mortality rate in patients with end-stage renal disease who are on dialysis has steadily fallen over the past 20 years, from an annual rate of about 25% in 1996 to 17% in 2014, suggesting that care improved during that time. Patients with transplants have a much lower mortality rate: less than 5% annually.⁵ But these data also highlight the persistent risk faced by patients with chronic kidney disease; even those with transplants have death rates comparable to those of patients with cancer, diabetes, or heart failure.

Death rates correlate with GFR

After the publication of definitions and staging by the National Kidney Foundation in 2002, Go et al⁶ studied more than 1 million patients with chronic kidney disease from the Kaiser Permanente Renal Registry and found that the rates of cardiovascular events and death from any cause increased with decreasing estimated GFR. These findings were confirmed in a later meta-analysis, which also found that an elevated urinary albumin-to-creatinine ratio (> 1.1 mg/mmol) is an independent predictor of all-cause mortality and cardiovascular mortality.⁷

Keith et al⁸ followed nearly 28,000 patients with chronic kidney disease (with an estimated GFR of less than 90 mL/min/1.73 m²) over 5 years. Patients with stage 3 disease (moderate disease, GFR = 30–59 mL/min/1.73 m²) were 20 times more likely to die than to progress to end-stage renal disease (24.3% vs 1.2%). Even those with stage 4 disease (severe disease, GFR = 15–29 mL/min/1.73 m²) were more than twice as likely to die as to progress to dialysis (45.7% vs 19.9%).

 

 

Heart disease risk increases with declining kidney function

Similar observations were made by Thompson et al¹⁰ based on the Alberta Kidney Disease Network database. They tracked cardiovascular causes of death and found that regardless of estimated GFR, cardiovascular deaths were most often attributed to ischemic heart disease (about 55%). Other trends were also apparent: as the GFR fell, the incidence of stroke decreased, and heart failure and valvular heart disease increased.

AFRICAN AMERICANS WITH KIDNEY DISEASE: A DISTINCT GROUP

African Americans constitute about 12% of the US population but account for:

• 31% of end-stage renal disease
• 34% of the kidney transplant waiting list
• 28% of kidney transplants in 2015 (12% of living donor transplants, 35% of deceased donor transplants).

In addition, African Americans with chronic kidney disease tend to be:

• Younger and have more advanced kidney disease than whites¹¹
• Much more likely than whites to have diabetes, and somewhat more likely to have hypertension

• 

  Adapted from Navaneethan SD, Schold JD, Arrigain S, Jolly SE, Nally JV Jr. Cause-specific deaths in non-dialysis-dependent CKD. J Am Soc Nephrol 2015; 26:2512–2520.

  More likely than whites to die of cardiovascular disease (37.4% vs 34.2%) (Figure 2).⁹

Overall, the prevalence of chronic kidney disease is slightly higher in African Americans than in whites. Interestingly, African Americans are slightly less likely than whites to have low estimated GFR values (6.2% vs 7.6% incidence of < 60 mL/min/1.73 m²) but are about 50% more likely to have proteinuria (12.3% vs 8.4% incidence of urine albumin-to-creatinine ratio ≥ 30 mg/g).

More likely to be on dialysis, but less likely to die

Although African Americans have only a slightly higher prevalence of chronic kidney disease (about 15% increased prevalence) than whites,¹² they are 3 times more likely to be on dialysis.

Nevertheless, for unknown reasons, African American adults on dialysis have about a 26% lower all-cause mortality rate than whites.⁵ One proposed explanation for this survival advantage has been that the mortality rate in African Americans with chronic kidney disease before entering dialysis is higher than in whites, leading to a “healthier population” on dialysis.¹³ However, this theory is based on a small study from more than a decade ago and has not been borne out by subsequent investigation.

African Americans with chronic kidney disease: Death rates not increased

African Americans over age 65 with chronic kidney disease have all-cause mortality rates similar to those of whites: about 11% annually. Breaking it down by disease severity, death rates in stage 3 disease are about 10% and jump to more than 15% in higher stages in both African Americans and whites.⁵

However, African Americans with chronic kidney disease have more heart disease and much more end-stage renal disease than whites.

Disease advances faster despite care

The incidence of end-stage renal disease is consistently more than 3 times higher in African Americans than in whites in the United States.^5,14

Multiple investigations have tried to determine why African Americans are disproportionately affected by progression of chronic kidney disease to end-stage renal disease. We recently examined this question in our Cleveland Clinic registry data. Even after adjusting for 17 variables (including demographics, comorbidities, insurance, medications, smoking, and chronic kidney disease stage), African Americans with chronic kidney disease were found to have an increased risk of progressing to end-stage renal disease compared with whites (subhazard ratio 1.38, 95% confidence interval 1.19–1.60).

We examined care measures from the Cleveland Clinic database. In terms of the number of laboratory tests ordered, clinic visits, and nephrology referrals, African Americans had at least as much care as whites, if not more. Similarly, African Americans’ access to renoprotective medicines (angiotensin-converting enzyme inhibitors, angiotensin II receptor blockers, statins, beta-blockers) was the same as or more than for whites.

Although the frequently attributed reasons surrounding compliance and socioeconomic issues are worthy of examination, they do not appear to completely explain the differences in incidence and outcomes. This dichotomy of a marginally increased prevalence of chronic kidney disease in African Americans with mortality rates similar to those of whites, yet with a 3 times higher incidence of end-stage renal disease in African Americans, suggests a faster progression of the disease in African Americans, which may be genetically based.

 

 

GENETIC VARIANTS FOUND

In 2010, two variant alleles of the APOL1 gene on chromosome 22 were found to be associated with nondiabetic kidney disease.¹⁵ Three nephropathies are associated with being homozygous for these alleles:

• Focal segmental glomerulosclerosis, the leading cause of nephrotic syndrome in African Americans
• Hypertension-associated kidney disease with scarring of glomeruli in vessels, the primary cause of end-stage renal disease in African Americans
• Human immunodeficiency virus (HIV)- associated nephropathy, usually a focal segmental glomerulosclerosis type of lesion.

The first two conditions are about 3 to 5 times more prevalent in African Americans than in whites, and HIV-associated nephropathy is about 20 to 30 times more common. 

African sleeping sickness and chronic kidney disease

Retrospective analysis of biologic samples from trials of kidney disease in African Americans has revealed interesting results.

From Parsa A, Kao WH, Xie D, et al; AASK Study Investigators; CRIC Study Investigators. APOL1 risk variants, race, and progression of chronic kidney disease. N Engl J Med 2013; 369:2183–2196. Reprinted with permission from Massachusetts Medical Society.

The Chronic Renal Insufficiency Cohort (CRIC) observation study¹⁸ enrolled patients with an estimated GFR of 20 to 70 mL/min/1.73 m², with a preference for African Americans and patients with diabetes. Nearly 3,000 participants had adequate samples for DNA testing. They found that African Americans with the double variant allele had worse outcomes, whether or not they had diabetes, compared with whites and African Americans without the homozygous gene variant.

Mechanism not well understood

The mechanism of renal injury is not well understood. Apolipoprotein L1, the protein coded for by APOL1, is a component of high-density lipoprotein. It is found in a different distribution pattern in people with normal kidneys vs those with nondiabetic kidney disease, especially in the arteries, arterioles, and podocytes.^19,20 It can be detected in blood plasma, but levels do not correlate with kidney disease.²¹ Not all patients with the high-risk variant develop chronic kidney disease; a “second hit” such as infection with HIV may be required.

Investigators have recently developed knockout mouse models of APOL1-associated kidney diseases that are helping to elucidate mechanisms.^22,23

EFFECT OF GENOTYPE ON KIDNEY TRANSPLANTS IN AFRICAN AMERICANS

African Americans receive about 30% of kidney transplants in the United States and represent about 15% to 20% of all donors.

Lee et al²⁴ reviewed 119 African American recipients of kidney transplants, about half of whom were homozygous for an APOL1 variant. After 5 years, no differences were found in allograft survival between recipients with 0, 1, or 2 risk alleles.

However, looking at the issue from the other side, Reeves-Daniel et al²⁵ studied the fate of more than 100 kidneys that were transplanted from African American donors, 16% of whom had the high-risk, homozygous genotype. In this case, graft failure was much likelier to occur with the high-risk donor kidneys (hazard ratio 3.84, P = .008). Similar outcomes were shown in a study of 2 centers²⁶ involving 675 transplants from deceased donors, 15% of which involved the high-risk genotype. The hazard ratio for graft failure was found to be 2.26 (P = .001) with high-risk donor kidneys.

These studies, which examined data from about 5 years after transplant, found that kidney failure does not tend to occur immediately in all cases, but gradually over time. Most high-risk kidneys were not lost within the 5 years of the studies.

The fact that the high-risk kidneys do not all fail immediately also suggests that a second hit is required for failure. Culprits postulated include a bacterial or viral infection (eg, BK virus, cytomegalovirus), ischemia or reperfusion injury, drug toxicity, and immune-mediated allograft injury (ie, rejection). 

 

 

Genetic testing advisable?

Genetic testing for APOL1 risk variants is on the horizon for kidney transplant. But at this point, providing guidance for patients can be tricky. Two case studies^27,28 and epidemiologic data suggest that donors homozygous for an APOL1 variant and those with a family history of end-stage kidney disease are at increased risk of chronic kidney disease. Even so, most recipients even of these high-risk organs have good outcomes. If an African American patient needs a kidney and his or her sibling offers one, it is difficult to advise against it when the evidence is weak for immediate risk and when other options may not be readily available. Further investigation is clearly needed into whether APOL1 variants and other biomarkers can predict an organ’s success as a transplant.

The National Institutes of Health are currently funding prospective longitudinal studies with the APOL1 Long-term Kidney Transplantation Outcomes Network (APOLLO) to determine the impact of APOL1 genetic factors on transplant recipients as well as on living donors. Possible second hits will also be studied, as will other markers of renal dysfunction or disease in donors. Researchers are actively investigating these important questions.

KEEPING SCIENCE RELEVANT

In a recent commentary related to the murine knockout model of APOL1-associated kidney disease, O’Toole et al offered insightful observations regarding the potential clinical impact of these new genetic discoveries.²³

As we study the genetics of kidney disease in African American patients, we should keep in mind 3 critical questions of clinical importance:

Will findings identify better treatments for chronic kidney disease? The AASK trial found that knowing the genetics did not affect outcomes of routine therapy. However, basic science investigations are currently underway targeting APOL1 variants which might reduce the increased kidney disease risk among people of African descent.

Should patients be genotyped for APOL1 risk variants? For patients with chronic kidney disease, it does not seem useful at this time. But for renal transplant donors, the answer is probably yes.

How does this discovery help us to understand our patients better? The implications are enormous for combatting the assumptions that rapid chronic kidney disease progression reflects poor patient compliance or other socioeconomic factors. We now understand that genetics, at least in part, drives renal disease outcomes in African American patients.

Editor’s note: This Medical Grand Rounds was presented as the 14th Annual Lawrence “Chris” Crain Memorial Lecture, a series that has been dedicated to discussing kidney disease, hypertension, and health care disparities in the African American community. In 1997, Dr. Crain became the first African American chief medical resident at Cleveland Clinic, and was a nephrology fellow in 1998–1999. Dr. Nally was his teacher and mentor.

African Americans have a greater burden of chronic kidney disease than whites. They are more than 3 times as likely as whites to develop end-stage renal disease, even after adjusting for age, disease stage, smoking, medications, and comorbidities. Why this is so has been the focus of much speculation and research.

This article reviews recent advances in the understanding of the progression of chronic kidney disease, with particular scrutiny of the disease in African Americans. Breakthroughs in genetics that help explain the greater disease burden in African Americans are also discussed, as well as implications for organ transplant screening.

ADVANCING UNDERSTANDING OF CHRONIC KIDNEY DISEASE

In the 1990s, dialysis rolls grew by 8% to 10% annually. Unfortunately, many patients first met with a nephrologist on the eve of their first dialysis treatment; there was not yet an adequate way to recognize the disease earlier and slow its progression. And disease definitions were not yet standardized, which led to inadequate metrics and hampered the ability to move disease management forward.

Standardizing definitions

The situation improved in 2002, when the National Kidney Foundation published clinical practice guidelines for chronic kidney disease that included disease definitions and staging.¹ Chronic kidney disease was defined as a structural or functional abnormality of the kidney lasting at least 3 months, as manifested by either of the following:

• Kidney damage, with or without decreased glomerular filtration rate (GFR), as defined by pathologic abnormalities or markers of kidney damage in the blood, urine, or on imaging tests

• 

  GFR less than 60 mL/min/1.73 m², with or without kidney damage.

A subsequent major advance was the recognition that not only GFR but also albuminuria was important for staging of chronic kidney disease (Figure 1).²

Developing large databases

Surveillance and monitoring of chronic kidney disease have generated large databases that enable researchers to detect trends in disease progression.

US Renal Data System. The US Renal Data System has collected and reported on data for more than 20 years from the National Health and Nutrition Examination Survey and the Centers for Medicare and Medicaid Services about chronic and end-stage kidney disease in the United States.

Cleveland Clinic database. Cleveland Clinic has developed a validated chronic kidney disease registry based on its electronic health record.³ The data include demographics (age, sex, ethnic group), comorbidities, medications, and complete laboratory data.⁴

Alberta Kidney Disease Network. This Canadian research consortium links large laboratory and demographic databases to facilitate defining patient populations, such as those with kidney disease and other comorbidities.

Kaiser Permanente Renal Registry. Kaiser Permanente of Northern California insures more than one-third of adults in the San Francisco Bay Area. The renal registry includes all adults whose kidney function is known. Data on age, sex, and racial or ethnic group are available from the health-plan databases.

DEATHS FROM KIDNEY DISEASE

The mortality rate in patients with end-stage renal disease who are on dialysis has steadily fallen over the past 20 years, from an annual rate of about 25% in 1996 to 17% in 2014, suggesting that care improved during that time. Patients with transplants have a much lower mortality rate: less than 5% annually.⁵ But these data also highlight the persistent risk faced by patients with chronic kidney disease; even those with transplants have death rates comparable to those of patients with cancer, diabetes, or heart failure.

Death rates correlate with GFR

After the publication of definitions and staging by the National Kidney Foundation in 2002, Go et al⁶ studied more than 1 million patients with chronic kidney disease from the Kaiser Permanente Renal Registry and found that the rates of cardiovascular events and death from any cause increased with decreasing estimated GFR. These findings were confirmed in a later meta-analysis, which also found that an elevated urinary albumin-to-creatinine ratio (> 1.1 mg/mmol) is an independent predictor of all-cause mortality and cardiovascular mortality.⁷

Keith et al⁸ followed nearly 28,000 patients with chronic kidney disease (with an estimated GFR of less than 90 mL/min/1.73 m²) over 5 years. Patients with stage 3 disease (moderate disease, GFR = 30–59 mL/min/1.73 m²) were 20 times more likely to die than to progress to end-stage renal disease (24.3% vs 1.2%). Even those with stage 4 disease (severe disease, GFR = 15–29 mL/min/1.73 m²) were more than twice as likely to die as to progress to dialysis (45.7% vs 19.9%).

 

 

Heart disease risk increases with declining kidney function

Similar observations were made by Thompson et al¹⁰ based on the Alberta Kidney Disease Network database. They tracked cardiovascular causes of death and found that regardless of estimated GFR, cardiovascular deaths were most often attributed to ischemic heart disease (about 55%). Other trends were also apparent: as the GFR fell, the incidence of stroke decreased, and heart failure and valvular heart disease increased.

AFRICAN AMERICANS WITH KIDNEY DISEASE: A DISTINCT GROUP

African Americans constitute about 12% of the US population but account for:

• 31% of end-stage renal disease
• 34% of the kidney transplant waiting list
• 28% of kidney transplants in 2015 (12% of living donor transplants, 35% of deceased donor transplants).

In addition, African Americans with chronic kidney disease tend to be:

• Younger and have more advanced kidney disease than whites¹¹
• Much more likely than whites to have diabetes, and somewhat more likely to have hypertension

• 

  Adapted from Navaneethan SD, Schold JD, Arrigain S, Jolly SE, Nally JV Jr. Cause-specific deaths in non-dialysis-dependent CKD. J Am Soc Nephrol 2015; 26:2512–2520.

  More likely than whites to die of cardiovascular disease (37.4% vs 34.2%) (Figure 2).⁹

Overall, the prevalence of chronic kidney disease is slightly higher in African Americans than in whites. Interestingly, African Americans are slightly less likely than whites to have low estimated GFR values (6.2% vs 7.6% incidence of < 60 mL/min/1.73 m²) but are about 50% more likely to have proteinuria (12.3% vs 8.4% incidence of urine albumin-to-creatinine ratio ≥ 30 mg/g).

More likely to be on dialysis, but less likely to die

Although African Americans have only a slightly higher prevalence of chronic kidney disease (about 15% increased prevalence) than whites,¹² they are 3 times more likely to be on dialysis.

Nevertheless, for unknown reasons, African American adults on dialysis have about a 26% lower all-cause mortality rate than whites.⁵ One proposed explanation for this survival advantage has been that the mortality rate in African Americans with chronic kidney disease before entering dialysis is higher than in whites, leading to a “healthier population” on dialysis.¹³ However, this theory is based on a small study from more than a decade ago and has not been borne out by subsequent investigation.

African Americans with chronic kidney disease: Death rates not increased

African Americans over age 65 with chronic kidney disease have all-cause mortality rates similar to those of whites: about 11% annually. Breaking it down by disease severity, death rates in stage 3 disease are about 10% and jump to more than 15% in higher stages in both African Americans and whites.⁵

However, African Americans with chronic kidney disease have more heart disease and much more end-stage renal disease than whites.

Disease advances faster despite care

The incidence of end-stage renal disease is consistently more than 3 times higher in African Americans than in whites in the United States.^5,14

Multiple investigations have tried to determine why African Americans are disproportionately affected by progression of chronic kidney disease to end-stage renal disease. We recently examined this question in our Cleveland Clinic registry data. Even after adjusting for 17 variables (including demographics, comorbidities, insurance, medications, smoking, and chronic kidney disease stage), African Americans with chronic kidney disease were found to have an increased risk of progressing to end-stage renal disease compared with whites (subhazard ratio 1.38, 95% confidence interval 1.19–1.60).

We examined care measures from the Cleveland Clinic database. In terms of the number of laboratory tests ordered, clinic visits, and nephrology referrals, African Americans had at least as much care as whites, if not more. Similarly, African Americans’ access to renoprotective medicines (angiotensin-converting enzyme inhibitors, angiotensin II receptor blockers, statins, beta-blockers) was the same as or more than for whites.

Although the frequently attributed reasons surrounding compliance and socioeconomic issues are worthy of examination, they do not appear to completely explain the differences in incidence and outcomes. This dichotomy of a marginally increased prevalence of chronic kidney disease in African Americans with mortality rates similar to those of whites, yet with a 3 times higher incidence of end-stage renal disease in African Americans, suggests a faster progression of the disease in African Americans, which may be genetically based.

 

 

GENETIC VARIANTS FOUND

In 2010, two variant alleles of the APOL1 gene on chromosome 22 were found to be associated with nondiabetic kidney disease.¹⁵ Three nephropathies are associated with being homozygous for these alleles:

• Focal segmental glomerulosclerosis, the leading cause of nephrotic syndrome in African Americans
• Hypertension-associated kidney disease with scarring of glomeruli in vessels, the primary cause of end-stage renal disease in African Americans
• Human immunodeficiency virus (HIV)- associated nephropathy, usually a focal segmental glomerulosclerosis type of lesion.

The first two conditions are about 3 to 5 times more prevalent in African Americans than in whites, and HIV-associated nephropathy is about 20 to 30 times more common. 

African sleeping sickness and chronic kidney disease

Retrospective analysis of biologic samples from trials of kidney disease in African Americans has revealed interesting results.

From Parsa A, Kao WH, Xie D, et al; AASK Study Investigators; CRIC Study Investigators. APOL1 risk variants, race, and progression of chronic kidney disease. N Engl J Med 2013; 369:2183–2196. Reprinted with permission from Massachusetts Medical Society.

The Chronic Renal Insufficiency Cohort (CRIC) observation study¹⁸ enrolled patients with an estimated GFR of 20 to 70 mL/min/1.73 m², with a preference for African Americans and patients with diabetes. Nearly 3,000 participants had adequate samples for DNA testing. They found that African Americans with the double variant allele had worse outcomes, whether or not they had diabetes, compared with whites and African Americans without the homozygous gene variant.

Mechanism not well understood

The mechanism of renal injury is not well understood. Apolipoprotein L1, the protein coded for by APOL1, is a component of high-density lipoprotein. It is found in a different distribution pattern in people with normal kidneys vs those with nondiabetic kidney disease, especially in the arteries, arterioles, and podocytes.^19,20 It can be detected in blood plasma, but levels do not correlate with kidney disease.²¹ Not all patients with the high-risk variant develop chronic kidney disease; a “second hit” such as infection with HIV may be required.

Investigators have recently developed knockout mouse models of APOL1-associated kidney diseases that are helping to elucidate mechanisms.^22,23

EFFECT OF GENOTYPE ON KIDNEY TRANSPLANTS IN AFRICAN AMERICANS

African Americans receive about 30% of kidney transplants in the United States and represent about 15% to 20% of all donors.

Lee et al²⁴ reviewed 119 African American recipients of kidney transplants, about half of whom were homozygous for an APOL1 variant. After 5 years, no differences were found in allograft survival between recipients with 0, 1, or 2 risk alleles.

However, looking at the issue from the other side, Reeves-Daniel et al²⁵ studied the fate of more than 100 kidneys that were transplanted from African American donors, 16% of whom had the high-risk, homozygous genotype. In this case, graft failure was much likelier to occur with the high-risk donor kidneys (hazard ratio 3.84, P = .008). Similar outcomes were shown in a study of 2 centers²⁶ involving 675 transplants from deceased donors, 15% of which involved the high-risk genotype. The hazard ratio for graft failure was found to be 2.26 (P = .001) with high-risk donor kidneys.

These studies, which examined data from about 5 years after transplant, found that kidney failure does not tend to occur immediately in all cases, but gradually over time. Most high-risk kidneys were not lost within the 5 years of the studies.

The fact that the high-risk kidneys do not all fail immediately also suggests that a second hit is required for failure. Culprits postulated include a bacterial or viral infection (eg, BK virus, cytomegalovirus), ischemia or reperfusion injury, drug toxicity, and immune-mediated allograft injury (ie, rejection). 

 

 

Genetic testing advisable?

Genetic testing for APOL1 risk variants is on the horizon for kidney transplant. But at this point, providing guidance for patients can be tricky. Two case studies^27,28 and epidemiologic data suggest that donors homozygous for an APOL1 variant and those with a family history of end-stage kidney disease are at increased risk of chronic kidney disease. Even so, most recipients even of these high-risk organs have good outcomes. If an African American patient needs a kidney and his or her sibling offers one, it is difficult to advise against it when the evidence is weak for immediate risk and when other options may not be readily available. Further investigation is clearly needed into whether APOL1 variants and other biomarkers can predict an organ’s success as a transplant.

The National Institutes of Health are currently funding prospective longitudinal studies with the APOL1 Long-term Kidney Transplantation Outcomes Network (APOLLO) to determine the impact of APOL1 genetic factors on transplant recipients as well as on living donors. Possible second hits will also be studied, as will other markers of renal dysfunction or disease in donors. Researchers are actively investigating these important questions.

KEEPING SCIENCE RELEVANT

In a recent commentary related to the murine knockout model of APOL1-associated kidney disease, O’Toole et al offered insightful observations regarding the potential clinical impact of these new genetic discoveries.²³

As we study the genetics of kidney disease in African American patients, we should keep in mind 3 critical questions of clinical importance:

Will findings identify better treatments for chronic kidney disease? The AASK trial found that knowing the genetics did not affect outcomes of routine therapy. However, basic science investigations are currently underway targeting APOL1 variants which might reduce the increased kidney disease risk among people of African descent.

Should patients be genotyped for APOL1 risk variants? For patients with chronic kidney disease, it does not seem useful at this time. But for renal transplant donors, the answer is probably yes.

How does this discovery help us to understand our patients better? The implications are enormous for combatting the assumptions that rapid chronic kidney disease progression reflects poor patient compliance or other socioeconomic factors. We now understand that genetics, at least in part, drives renal disease outcomes in African American patients.

Editor’s note: This Medical Grand Rounds was presented as the 14th Annual Lawrence “Chris” Crain Memorial Lecture, a series that has been dedicated to discussing kidney disease, hypertension, and health care disparities in the African American community. In 1997, Dr. Crain became the first African American chief medical resident at Cleveland Clinic, and was a nephrology fellow in 1998–1999. Dr. Nally was his teacher and mentor.

African Americans have a greater burden of chronic kidney disease than whites. They are more than 3 times as likely as whites to develop end-stage renal disease, even after adjusting for age, disease stage, smoking, medications, and comorbidities. Why this is so has been the focus of much speculation and research.

This article reviews recent advances in the understanding of the progression of chronic kidney disease, with particular scrutiny of the disease in African Americans. Breakthroughs in genetics that help explain the greater disease burden in African Americans are also discussed, as well as implications for organ transplant screening.

ADVANCING UNDERSTANDING OF CHRONIC KIDNEY DISEASE

In the 1990s, dialysis rolls grew by 8% to 10% annually. Unfortunately, many patients first met with a nephrologist on the eve of their first dialysis treatment; there was not yet an adequate way to recognize the disease earlier and slow its progression. And disease definitions were not yet standardized, which led to inadequate metrics and hampered the ability to move disease management forward.

Standardizing definitions

The situation improved in 2002, when the National Kidney Foundation published clinical practice guidelines for chronic kidney disease that included disease definitions and staging.¹ Chronic kidney disease was defined as a structural or functional abnormality of the kidney lasting at least 3 months, as manifested by either of the following:

• Kidney damage, with or without decreased glomerular filtration rate (GFR), as defined by pathologic abnormalities or markers of kidney damage in the blood, urine, or on imaging tests

• 

  GFR less than 60 mL/min/1.73 m², with or without kidney damage.

A subsequent major advance was the recognition that not only GFR but also albuminuria was important for staging of chronic kidney disease (Figure 1).²

Developing large databases

Surveillance and monitoring of chronic kidney disease have generated large databases that enable researchers to detect trends in disease progression.

US Renal Data System. The US Renal Data System has collected and reported on data for more than 20 years from the National Health and Nutrition Examination Survey and the Centers for Medicare and Medicaid Services about chronic and end-stage kidney disease in the United States.

Cleveland Clinic database. Cleveland Clinic has developed a validated chronic kidney disease registry based on its electronic health record.³ The data include demographics (age, sex, ethnic group), comorbidities, medications, and complete laboratory data.⁴

Alberta Kidney Disease Network. This Canadian research consortium links large laboratory and demographic databases to facilitate defining patient populations, such as those with kidney disease and other comorbidities.

Kaiser Permanente Renal Registry. Kaiser Permanente of Northern California insures more than one-third of adults in the San Francisco Bay Area. The renal registry includes all adults whose kidney function is known. Data on age, sex, and racial or ethnic group are available from the health-plan databases.

DEATHS FROM KIDNEY DISEASE

The mortality rate in patients with end-stage renal disease who are on dialysis has steadily fallen over the past 20 years, from an annual rate of about 25% in 1996 to 17% in 2014, suggesting that care improved during that time. Patients with transplants have a much lower mortality rate: less than 5% annually.⁵ But these data also highlight the persistent risk faced by patients with chronic kidney disease; even those with transplants have death rates comparable to those of patients with cancer, diabetes, or heart failure.

Death rates correlate with GFR

After the publication of definitions and staging by the National Kidney Foundation in 2002, Go et al⁶ studied more than 1 million patients with chronic kidney disease from the Kaiser Permanente Renal Registry and found that the rates of cardiovascular events and death from any cause increased with decreasing estimated GFR. These findings were confirmed in a later meta-analysis, which also found that an elevated urinary albumin-to-creatinine ratio (> 1.1 mg/mmol) is an independent predictor of all-cause mortality and cardiovascular mortality.⁷

Keith et al⁸ followed nearly 28,000 patients with chronic kidney disease (with an estimated GFR of less than 90 mL/min/1.73 m²) over 5 years. Patients with stage 3 disease (moderate disease, GFR = 30–59 mL/min/1.73 m²) were 20 times more likely to die than to progress to end-stage renal disease (24.3% vs 1.2%). Even those with stage 4 disease (severe disease, GFR = 15–29 mL/min/1.73 m²) were more than twice as likely to die as to progress to dialysis (45.7% vs 19.9%).

 

 

Heart disease risk increases with declining kidney function

Similar observations were made by Thompson et al¹⁰ based on the Alberta Kidney Disease Network database. They tracked cardiovascular causes of death and found that regardless of estimated GFR, cardiovascular deaths were most often attributed to ischemic heart disease (about 55%). Other trends were also apparent: as the GFR fell, the incidence of stroke decreased, and heart failure and valvular heart disease increased.

AFRICAN AMERICANS WITH KIDNEY DISEASE: A DISTINCT GROUP

African Americans constitute about 12% of the US population but account for:

• 31% of end-stage renal disease
• 34% of the kidney transplant waiting list
• 28% of kidney transplants in 2015 (12% of living donor transplants, 35% of deceased donor transplants).

In addition, African Americans with chronic kidney disease tend to be:

• Younger and have more advanced kidney disease than whites¹¹
• Much more likely than whites to have diabetes, and somewhat more likely to have hypertension

• 

  Adapted from Navaneethan SD, Schold JD, Arrigain S, Jolly SE, Nally JV Jr. Cause-specific deaths in non-dialysis-dependent CKD. J Am Soc Nephrol 2015; 26:2512–2520.

  More likely than whites to die of cardiovascular disease (37.4% vs 34.2%) (Figure 2).⁹

Overall, the prevalence of chronic kidney disease is slightly higher in African Americans than in whites. Interestingly, African Americans are slightly less likely than whites to have low estimated GFR values (6.2% vs 7.6% incidence of < 60 mL/min/1.73 m²) but are about 50% more likely to have proteinuria (12.3% vs 8.4% incidence of urine albumin-to-creatinine ratio ≥ 30 mg/g).

More likely to be on dialysis, but less likely to die

Although African Americans have only a slightly higher prevalence of chronic kidney disease (about 15% increased prevalence) than whites,¹² they are 3 times more likely to be on dialysis.

Nevertheless, for unknown reasons, African American adults on dialysis have about a 26% lower all-cause mortality rate than whites.⁵ One proposed explanation for this survival advantage has been that the mortality rate in African Americans with chronic kidney disease before entering dialysis is higher than in whites, leading to a “healthier population” on dialysis.¹³ However, this theory is based on a small study from more than a decade ago and has not been borne out by subsequent investigation.

African Americans with chronic kidney disease: Death rates not increased

African Americans over age 65 with chronic kidney disease have all-cause mortality rates similar to those of whites: about 11% annually. Breaking it down by disease severity, death rates in stage 3 disease are about 10% and jump to more than 15% in higher stages in both African Americans and whites.⁵

However, African Americans with chronic kidney disease have more heart disease and much more end-stage renal disease than whites.

Disease advances faster despite care

The incidence of end-stage renal disease is consistently more than 3 times higher in African Americans than in whites in the United States.^5,14

Multiple investigations have tried to determine why African Americans are disproportionately affected by progression of chronic kidney disease to end-stage renal disease. We recently examined this question in our Cleveland Clinic registry data. Even after adjusting for 17 variables (including demographics, comorbidities, insurance, medications, smoking, and chronic kidney disease stage), African Americans with chronic kidney disease were found to have an increased risk of progressing to end-stage renal disease compared with whites (subhazard ratio 1.38, 95% confidence interval 1.19–1.60).

We examined care measures from the Cleveland Clinic database. In terms of the number of laboratory tests ordered, clinic visits, and nephrology referrals, African Americans had at least as much care as whites, if not more. Similarly, African Americans’ access to renoprotective medicines (angiotensin-converting enzyme inhibitors, angiotensin II receptor blockers, statins, beta-blockers) was the same as or more than for whites.

Although the frequently attributed reasons surrounding compliance and socioeconomic issues are worthy of examination, they do not appear to completely explain the differences in incidence and outcomes. This dichotomy of a marginally increased prevalence of chronic kidney disease in African Americans with mortality rates similar to those of whites, yet with a 3 times higher incidence of end-stage renal disease in African Americans, suggests a faster progression of the disease in African Americans, which may be genetically based.

 

 

GENETIC VARIANTS FOUND

In 2010, two variant alleles of the APOL1 gene on chromosome 22 were found to be associated with nondiabetic kidney disease.¹⁵ Three nephropathies are associated with being homozygous for these alleles:

• Focal segmental glomerulosclerosis, the leading cause of nephrotic syndrome in African Americans
• Hypertension-associated kidney disease with scarring of glomeruli in vessels, the primary cause of end-stage renal disease in African Americans
• Human immunodeficiency virus (HIV)- associated nephropathy, usually a focal segmental glomerulosclerosis type of lesion.

The first two conditions are about 3 to 5 times more prevalent in African Americans than in whites, and HIV-associated nephropathy is about 20 to 30 times more common. 

African sleeping sickness and chronic kidney disease

Retrospective analysis of biologic samples from trials of kidney disease in African Americans has revealed interesting results.

From Parsa A, Kao WH, Xie D, et al; AASK Study Investigators; CRIC Study Investigators. APOL1 risk variants, race, and progression of chronic kidney disease. N Engl J Med 2013; 369:2183–2196. Reprinted with permission from Massachusetts Medical Society.

The Chronic Renal Insufficiency Cohort (CRIC) observation study¹⁸ enrolled patients with an estimated GFR of 20 to 70 mL/min/1.73 m², with a preference for African Americans and patients with diabetes. Nearly 3,000 participants had adequate samples for DNA testing. They found that African Americans with the double variant allele had worse outcomes, whether or not they had diabetes, compared with whites and African Americans without the homozygous gene variant.

Mechanism not well understood

The mechanism of renal injury is not well understood. Apolipoprotein L1, the protein coded for by APOL1, is a component of high-density lipoprotein. It is found in a different distribution pattern in people with normal kidneys vs those with nondiabetic kidney disease, especially in the arteries, arterioles, and podocytes.^19,20 It can be detected in blood plasma, but levels do not correlate with kidney disease.²¹ Not all patients with the high-risk variant develop chronic kidney disease; a “second hit” such as infection with HIV may be required.

Investigators have recently developed knockout mouse models of APOL1-associated kidney diseases that are helping to elucidate mechanisms.^22,23

EFFECT OF GENOTYPE ON KIDNEY TRANSPLANTS IN AFRICAN AMERICANS

African Americans receive about 30% of kidney transplants in the United States and represent about 15% to 20% of all donors.

Lee et al²⁴ reviewed 119 African American recipients of kidney transplants, about half of whom were homozygous for an APOL1 variant. After 5 years, no differences were found in allograft survival between recipients with 0, 1, or 2 risk alleles.

However, looking at the issue from the other side, Reeves-Daniel et al²⁵ studied the fate of more than 100 kidneys that were transplanted from African American donors, 16% of whom had the high-risk, homozygous genotype. In this case, graft failure was much likelier to occur with the high-risk donor kidneys (hazard ratio 3.84, P = .008). Similar outcomes were shown in a study of 2 centers²⁶ involving 675 transplants from deceased donors, 15% of which involved the high-risk genotype. The hazard ratio for graft failure was found to be 2.26 (P = .001) with high-risk donor kidneys.

These studies, which examined data from about 5 years after transplant, found that kidney failure does not tend to occur immediately in all cases, but gradually over time. Most high-risk kidneys were not lost within the 5 years of the studies.

The fact that the high-risk kidneys do not all fail immediately also suggests that a second hit is required for failure. Culprits postulated include a bacterial or viral infection (eg, BK virus, cytomegalovirus), ischemia or reperfusion injury, drug toxicity, and immune-mediated allograft injury (ie, rejection). 

 

 

Genetic testing advisable?

Genetic testing for APOL1 risk variants is on the horizon for kidney transplant. But at this point, providing guidance for patients can be tricky. Two case studies^27,28 and epidemiologic data suggest that donors homozygous for an APOL1 variant and those with a family history of end-stage kidney disease are at increased risk of chronic kidney disease. Even so, most recipients even of these high-risk organs have good outcomes. If an African American patient needs a kidney and his or her sibling offers one, it is difficult to advise against it when the evidence is weak for immediate risk and when other options may not be readily available. Further investigation is clearly needed into whether APOL1 variants and other biomarkers can predict an organ’s success as a transplant.

The National Institutes of Health are currently funding prospective longitudinal studies with the APOL1 Long-term Kidney Transplantation Outcomes Network (APOLLO) to determine the impact of APOL1 genetic factors on transplant recipients as well as on living donors. Possible second hits will also be studied, as will other markers of renal dysfunction or disease in donors. Researchers are actively investigating these important questions.

KEEPING SCIENCE RELEVANT

In a recent commentary related to the murine knockout model of APOL1-associated kidney disease, O’Toole et al offered insightful observations regarding the potential clinical impact of these new genetic discoveries.²³

As we study the genetics of kidney disease in African American patients, we should keep in mind 3 critical questions of clinical importance:

Will findings identify better treatments for chronic kidney disease? The AASK trial found that knowing the genetics did not affect outcomes of routine therapy. However, basic science investigations are currently underway targeting APOL1 variants which might reduce the increased kidney disease risk among people of African descent.

Should patients be genotyped for APOL1 risk variants? For patients with chronic kidney disease, it does not seem useful at this time. But for renal transplant donors, the answer is probably yes.

How does this discovery help us to understand our patients better? The implications are enormous for combatting the assumptions that rapid chronic kidney disease progression reflects poor patient compliance or other socioeconomic factors. We now understand that genetics, at least in part, drives renal disease outcomes in African American patients.

Click here to access the November 2017 Digital Edition.

Table of Contents

• HIV-Positive Veteran With Progressive Visual Changes
• Duodenal Perforation After Endoscopic Procedure
• Providers’ Attitudes and Knowledge of Lesbian, Gay, Bisexual, and Transgender Health
• Shoulder Pain—Is It From the Shoulder, Neck, or Both?
• Responding to the Opiod Crisis: An IHS Pharmacist-Led Pain Management Clinic
• Introducing the VA Boston Medical Forum
• Psychological Consequences of Detainee Operations
• Bryant Homer Womack: Honoring Sacrifice

Click here to access the November 2017 Digital Edition.

Table of Contents

• HIV-Positive Veteran With Progressive Visual Changes
• Duodenal Perforation After Endoscopic Procedure
• Providers’ Attitudes and Knowledge of Lesbian, Gay, Bisexual, and Transgender Health
• Shoulder Pain—Is It From the Shoulder, Neck, or Both?
• Responding to the Opiod Crisis: An IHS Pharmacist-Led Pain Management Clinic
• Introducing the VA Boston Medical Forum
• Psychological Consequences of Detainee Operations
• Bryant Homer Womack: Honoring Sacrifice

Click here to access the November 2017 Digital Edition.

Table of Contents

• HIV-Positive Veteran With Progressive Visual Changes
• Duodenal Perforation After Endoscopic Procedure
• Providers’ Attitudes and Knowledge of Lesbian, Gay, Bisexual, and Transgender Health
• Shoulder Pain—Is It From the Shoulder, Neck, or Both?
• Responding to the Opiod Crisis: An IHS Pharmacist-Led Pain Management Clinic
• Introducing the VA Boston Medical Forum
• Psychological Consequences of Detainee Operations
• Bryant Homer Womack: Honoring Sacrifice

►Lakshmana Swamy, MD, chief medical resident, VA Boston Healthcare System (VABHS) and Boston Medical Center. Dr. Serrao, when you hear about vision changes in a patient with HIV, what differential diagnosis is generated? What epidemiologic or historical factors can help distinguish these entities?

►Richard Serrao, MD, Infectious Disease Service, VABHS and assistant professor of medicine, Boston University School of Medicine. The differential diagnoses for vision changes in a patient with HIV is based on the overall immunosuppression of the patient: the lower the patient’s CD4 count, the higher the number of etiologies.¹ The portions of the visual pathway as well as the pattern of vision loss are useful in narrowing the differential. For example, monocular visual disturbances with dermatomal vesicles within the ophthalmic division of the trigeminal nerve strongly implicates varicella zoster retinitis or keratitis; abducens nerve palsy could suggest granulomatous basilar meningitis from cryptococcosis. Likewise, ongoing fevers in an advanced AIDS patient with concomitant colitis, hepatitis, and pneumonitis is strongly suspicious for cytomegalovirus (CMV) retinitis with wide dissemination.

Geographic epidemiologic factors can suggest pathogens more prevalent to certain regions of the world, such as histoplasma chorioretinitis in a resident of the central and eastern U.S. or tuberculosis in a returning traveler. Likewise, a cat owner or one who consumes steak tartare increases the likelihood for toxoplasma retinochoroiditis, or syphilis in men who have sex with men (MSM) in the U.S. given that the majority of new cases occur in this patient population. Other clues one should consider include the presence of splinter hemorrhages in the extremities in an intravenous drug user, raising the possibility of embolic endophthalmitis from bacterial or fungal endocarditis. A variety of other diagnoses can certainly occur as a result of drug treatment (uveitis from rifampin, for example), immune reconstitution from HAART, infections with other HIV-associated pathogens, such as Pneumocystis jiroveci, and many non-HIV-related ocular diseases.

►Dr. Swamy. Dr. Butler, what concerns do you have when you hear about an HIV-infected patient with vision loss from the ophthalmology perspective?

►Nicholas Butler, MD, Ophthalmology Service, Uveitis and Ocular Immunology, VABHS and assistant professor of ophthalmology, Harvard Medical School. Of course, patients with HIV suffer from common causes of vision loss—cataract, glaucoma, diabetes, macular degeneration, for instance—just like those without HIV infection. If there is no significant immunodeficiency, then the patient’s HIV status would be less relevant, and these more common causes of vision loss should be pursued. My first task would be to determine the patient’s most recent CD4 T-cell count.

Assuming an HIV-positive individual is experiencing visual symptoms related to his/her underlying HIV infection (especially in the setting of CD4 counts < 200 cells/mm³), ocular opportunistic infections (OOI) come to mind first. Despite a reduction in incidence of 75% to 80% in the HAART-era, CMV retinitis remains the most common OOI in patients with AIDS and carries the greatest risk of ocular morbidity.² In fact, based on enrollment data for the Longitudinal Study of the Ocular Complications of AIDS (LSOCA), the prevalence of CMV retinitis among patients with AIDS is more than 20-fold higher than all other ocular complications of AIDS (OOIs and ocular neoplastic disease), including Kaposi sarcoma, lymphoma, herpes zoster ophthalmicus, ocular syphilis, ocular toxoplasma, necrotizing herpetic retinitis, cryptococcal choroiditis, and pneumocystis choroiditis.³ Beyond ocular opportunistic infections, the most common retinal finding in HIV-positive people is HIV retinopathy, nonspecific microvascular findings in the retina affecting nearly 70% of those with advanced HIV disease. Fortunately, HIV retinopathy is generally asymptomatic.⁴

►Dr. Swamy. Thank you for those explanations. Based on Dr. Serrao’s differential, it is worth noting that this patient is MSM. He was evaluated in urgent care with the initial examination showing a temperature of 98.0° F, pulse 83 beats per minute, and blood pressure 110/70 mm Hg. The eye exam showed no injection with normal extraocular movements. Initial laboratory data were notable for a CD4 count of 730 cells/mm³ with fewer than 20 HIV viral copies/mL. Cytomegalovirus immunoglobulin G (IgG) was positive, and immunoglobulin M (IgM) was negative. A Lyme antibody was positive with negative IgM and IgG by Western blot. Additional tests can be seen in Tables 1 and 2. The patient has good immunologic and virologic control. How does this change your thinking about the case?

►Dr. Serrao. His CD4 count is well above 350, increasing the likelihood of a relatively uncomplicated course and treatment. Cytomegalovirus antibodies reflect prior infection. As CMV generally does not manifest with disease of any variety (including CMV retinitis) at this high CD4 count, one can presume he does not have CMV retinitis as a cause for his visual changes. CMV retinitis occurs mainly when substantial CD4 depletion has occurred (typically less than 50 cells/mm³). A positive Lyme antibody screen, not specific to Lyme, can be falsely positive in other treponema diseases (eg, Treponema pallidum, the etiologic organism of syphilis) as evidenced by negative confirmatory Western blot IgG and IgM. Antineutrophil cystoplasmic antibodies, lysozyme, angiotensin-converting enzyme, rapid plasma reagin (RPR), herpes simplex virus, toxoplasma are generally included in the workup for the differential of uveitis, retinitis, choroiditis, etc.

►Dr. Swamy. Based on the visual changes, the patient was referred for urgent ophthalmologic evaluation. Dr. Butler, when should a generalist consider urgent ophthalmology referral?

►Dr. Butler. In general, all patients with acute (and significant) vision loss should be referred immediately to an ophthalmologist. The challenge for the general practitioner is determining the true extent of the reported vision loss. If possible, some assessment of visual acuity should be obtained, testing each eye independently and with the correct glasses correction (ie, the patient’s distance glasses if the test object is 12 feet or more from the patient or their reading glasses if the test object is held inside arm’s length). If the general practitioner does not have access to an eye chart or near card, any assessment of vision with an appropriate description will be useful (eg, the patient can quickly count fingers at 15 feet in the unaffected eye, but the eye with reported vision loss cannot reliably count fingers outside of 2 feet). Additional ocular symptoms associated with the vision loss, such as pain, redness, photophobia, new flashes or floaters, increase the urgency of the referral. The threshold for referral for any ocular complaint is lower compared with that of the general population for those with evidence of immunodeficiency, such as for this patient with HIV. Any CD4 count < 200 cells/mm³ should raise the practitioner’s concern for an ocular opportunistic infection, with the greatest concern with CD4 counts < 50 cells/mm³.

►Dr. Swamy. The patient underwent further testing in the ophthalmology clinic. Dr. Butler, can you please interpret the funduscopic exam?

►Dr. Butler. Both eyes demonstrate findings (microaneurysms and small dot-blot hemorrhages) consistent with moderate nonproliferative diabetic retinopathy (Figure 1A, white arrows). HIV-associated retinopathy could produce similar findings, but it is not generally seen with CD4 counts > 200 cells/mm³. Additionally, in the left eye, there is a diffuse patch of retinal whitening (retinitis) associated with the inferotemporal vascular arcades (Figure 1B, white arrows). The entire area involved is poorly circumscribed and the whitening is subtle in areas. Overlying some areas of deeper, ground-glass whitening there are scattered, punctate white spots (Figure 1B, green arrows). Wickremasinghe and colleagues described this pattern of retinitis and suggested that it had a high positive-predictive value in the diagnosis of ocular syphilis.⁵

►Dr. Swamy. The patient then underwent fluorescein angiography and optical coherence tomography (OCT). Dr. Butler, what did the fluorescein angiography show?

►Dr. Butler. The fluorescein angiogram in both eyes revealed leakage of dye consistent with diabetic retinopathy, with the right eye (OD) worse than the left (OS). Additionally, the areas of active retinitis in the left eye displayed gradual staining with leopard-spot changes, along with late leakage of fluorescein dye, indicating vasculopathy in the infected area (Figure 2, arrows). The patient also underwent OCT in the left eye (images not displayed) demonstrating vitreous cells (vitritis), patches of inner retinal thickening with hyperreflectivity, and hyperreflective nodules at the level of the retinal pigment epithelium with overlying photoreceptor disruption. These OCT findings are fairly stereotypic for syphilitic chorioretinitis.⁶

►Dr. Swamy. Based on the ophthalmic findings, a diagnosis of ocular syphilis was made. Dr. Serrao, what should internists consider as they evaluate and manage a patient with ocular syphilis?

►Dr. Serrao. Although isolated ocular involvement from syphilis is possible, the majority of patients (up to 85%) with HIV can present with concomitant central nervous system infection and about 30% present with symptomatic neurosyphilis (a typical late manifestation of this disease) that reflects the aggressiveness, accelerated course and propensity for wide dissemination of syphilis in this patient population.⁷

The presence of concomitant cutaneous rashes should prompt universal precautions, because transmission can occur via skin to skin contact. Clinicians should watch for the Jarisch-Herxheimer reaction during treatment, a syndrome of fever, myalgias, and headache, which results from circulating cytokines produced because of rapidly dying spirochetes that could mimic a penicillin drug reaction, yet is treated supportively.

As syphilis is sexually acquired, clinicians should test for coexistent sexually transmitted infections, vaccinate for those that are preventable (eg, hepatitis B), notify sexual partners via assistance from local departments of public health, and assess for coexistent drug use and offer counseling in order to optimize risk reduction. Special attention should be paid to virologic control of HIV since some studies have shown an increase in the propensity for breakthrough HIV viremia while on effective ART.⁹ This should warrant counseling for ongoing optimal ART adherence and close monitoring in the follow-up visits with a provider specialized in the treatment of syphilis and HIV.

►Dr. Swamy. A lumbar puncture is performed with the results listed in Table 2. Dr. Serrao, is the CSF consistent with neurosyphilis? What would you do next?

►Dr. Serrao. The lumbar puncture is inflammatory with a lymphocytic predominance, consistent with active ocular/neurosyphilis. The CSF Venereal Disease Research Laboratory test is specific but not sensitive so a negative value does not rule out the presence of central nervous system infection.¹⁰ The CSF fluorescent treponemal antibody (CSF FTA-ABS) is sensitive but not specific. In this case, the ocular findings, positive serum RPR, CSF lymphocytic predominance, and CSF FTA ABS strongly supports the diagnosis of ocular/early neurosyphilis in a patient with HIV infection in whom early aggressive treatment is warranted to prevent rapid progression/potential loss of vision.¹¹

►Dr. Swamy. Dr. Butler, how does syphilis behave in the eye as compared to other infectious or inflammatory diseases? Do visual symptoms respond well to treatment?

►Dr. Butler. As opposed to the dramatic reduction in rates and severity of CMV retinitis, HAART has had a negligible effect on ocular syphilis in the setting of HIV coinfection; in fact, rates of syphilis, including ocular syphilis, are currently surging world-wide, and HIV coinfection portends a worse prognosis.¹² This is especially true among gay men. More so, there appears to be no correlation between CD4 count and incidence of developing ocular syphilis, as opposed to CMV retinitis, which occurs far more frequently in those with CD4 counts < 50 cells/mm³. In keeping with its epithet as one of the “Great Imitators,” syphilis can affect virtually every tissue of the eye—conjunctiva, sclera, cornea, iris, lens, vitreous, retina, choroid, optic nerve—unlike other OOI, such as CMV or toxoplasma, which generally hone to the retina. Nonetheless, various findings and patterns on clinical exam and ancillary testing, such as the more recently described punctate inner retinitis (as seen in our patient) and the more classic acute syphilitic posterior placoid chorioretinitis, carry high specificity for ocular syphilis.¹³

Patients with ocular syphilis should be treated according to neurosyphilis treatment protocols. In general, these patients respond very well to treatment with resolution of the ocular findings and recovery of complete, or nearly so, visual function, as long as an excessive delay between diagnosis and proper treatment does not occur.¹⁴

►Dr. Swamy. Following this testing, the patient completed 14 days of IV penicillin with resolution of symptoms. He had no further vision complaints. He was started on Triumeq (abacavir, dolutegravir, and lamivudine) with good adherence to therapy. Dr. Serrao, in 2016 the CDC released a clinical advisory about ocular syphilis. Can you tell us about why this is an important diagnosis to be aware of today?

►Dr. Serrao. As with any disease, the epidemiologic characteristics of an infection like syphilis allow the clinician to more carefully entertain such a diagnosis in any one individual by improving the index of suspicion for a particular disease. Awareness of an increase in ocular syphilis in HIV positive MSM allows for a more timely assessment and subsequent treatment with the goal of preventing loss of vision.¹⁵

►Lakshmana Swamy, MD, chief medical resident, VA Boston Healthcare System (VABHS) and Boston Medical Center. Dr. Serrao, when you hear about vision changes in a patient with HIV, what differential diagnosis is generated? What epidemiologic or historical factors can help distinguish these entities?

►Richard Serrao, MD, Infectious Disease Service, VABHS and assistant professor of medicine, Boston University School of Medicine. The differential diagnoses for vision changes in a patient with HIV is based on the overall immunosuppression of the patient: the lower the patient’s CD4 count, the higher the number of etiologies.¹ The portions of the visual pathway as well as the pattern of vision loss are useful in narrowing the differential. For example, monocular visual disturbances with dermatomal vesicles within the ophthalmic division of the trigeminal nerve strongly implicates varicella zoster retinitis or keratitis; abducens nerve palsy could suggest granulomatous basilar meningitis from cryptococcosis. Likewise, ongoing fevers in an advanced AIDS patient with concomitant colitis, hepatitis, and pneumonitis is strongly suspicious for cytomegalovirus (CMV) retinitis with wide dissemination.

Geographic epidemiologic factors can suggest pathogens more prevalent to certain regions of the world, such as histoplasma chorioretinitis in a resident of the central and eastern U.S. or tuberculosis in a returning traveler. Likewise, a cat owner or one who consumes steak tartare increases the likelihood for toxoplasma retinochoroiditis, or syphilis in men who have sex with men (MSM) in the U.S. given that the majority of new cases occur in this patient population. Other clues one should consider include the presence of splinter hemorrhages in the extremities in an intravenous drug user, raising the possibility of embolic endophthalmitis from bacterial or fungal endocarditis. A variety of other diagnoses can certainly occur as a result of drug treatment (uveitis from rifampin, for example), immune reconstitution from HAART, infections with other HIV-associated pathogens, such as Pneumocystis jiroveci, and many non-HIV-related ocular diseases.

►Dr. Swamy. Dr. Butler, what concerns do you have when you hear about an HIV-infected patient with vision loss from the ophthalmology perspective?

►Nicholas Butler, MD, Ophthalmology Service, Uveitis and Ocular Immunology, VABHS and assistant professor of ophthalmology, Harvard Medical School. Of course, patients with HIV suffer from common causes of vision loss—cataract, glaucoma, diabetes, macular degeneration, for instance—just like those without HIV infection. If there is no significant immunodeficiency, then the patient’s HIV status would be less relevant, and these more common causes of vision loss should be pursued. My first task would be to determine the patient’s most recent CD4 T-cell count.

Assuming an HIV-positive individual is experiencing visual symptoms related to his/her underlying HIV infection (especially in the setting of CD4 counts < 200 cells/mm³), ocular opportunistic infections (OOI) come to mind first. Despite a reduction in incidence of 75% to 80% in the HAART-era, CMV retinitis remains the most common OOI in patients with AIDS and carries the greatest risk of ocular morbidity.² In fact, based on enrollment data for the Longitudinal Study of the Ocular Complications of AIDS (LSOCA), the prevalence of CMV retinitis among patients with AIDS is more than 20-fold higher than all other ocular complications of AIDS (OOIs and ocular neoplastic disease), including Kaposi sarcoma, lymphoma, herpes zoster ophthalmicus, ocular syphilis, ocular toxoplasma, necrotizing herpetic retinitis, cryptococcal choroiditis, and pneumocystis choroiditis.³ Beyond ocular opportunistic infections, the most common retinal finding in HIV-positive people is HIV retinopathy, nonspecific microvascular findings in the retina affecting nearly 70% of those with advanced HIV disease. Fortunately, HIV retinopathy is generally asymptomatic.⁴

►Dr. Swamy. Thank you for those explanations. Based on Dr. Serrao’s differential, it is worth noting that this patient is MSM. He was evaluated in urgent care with the initial examination showing a temperature of 98.0° F, pulse 83 beats per minute, and blood pressure 110/70 mm Hg. The eye exam showed no injection with normal extraocular movements. Initial laboratory data were notable for a CD4 count of 730 cells/mm³ with fewer than 20 HIV viral copies/mL. Cytomegalovirus immunoglobulin G (IgG) was positive, and immunoglobulin M (IgM) was negative. A Lyme antibody was positive with negative IgM and IgG by Western blot. Additional tests can be seen in Tables 1 and 2. The patient has good immunologic and virologic control. How does this change your thinking about the case?

►Dr. Serrao. His CD4 count is well above 350, increasing the likelihood of a relatively uncomplicated course and treatment. Cytomegalovirus antibodies reflect prior infection. As CMV generally does not manifest with disease of any variety (including CMV retinitis) at this high CD4 count, one can presume he does not have CMV retinitis as a cause for his visual changes. CMV retinitis occurs mainly when substantial CD4 depletion has occurred (typically less than 50 cells/mm³). A positive Lyme antibody screen, not specific to Lyme, can be falsely positive in other treponema diseases (eg, Treponema pallidum, the etiologic organism of syphilis) as evidenced by negative confirmatory Western blot IgG and IgM. Antineutrophil cystoplasmic antibodies, lysozyme, angiotensin-converting enzyme, rapid plasma reagin (RPR), herpes simplex virus, toxoplasma are generally included in the workup for the differential of uveitis, retinitis, choroiditis, etc.

►Dr. Swamy. Based on the visual changes, the patient was referred for urgent ophthalmologic evaluation. Dr. Butler, when should a generalist consider urgent ophthalmology referral?

►Dr. Butler. In general, all patients with acute (and significant) vision loss should be referred immediately to an ophthalmologist. The challenge for the general practitioner is determining the true extent of the reported vision loss. If possible, some assessment of visual acuity should be obtained, testing each eye independently and with the correct glasses correction (ie, the patient’s distance glasses if the test object is 12 feet or more from the patient or their reading glasses if the test object is held inside arm’s length). If the general practitioner does not have access to an eye chart or near card, any assessment of vision with an appropriate description will be useful (eg, the patient can quickly count fingers at 15 feet in the unaffected eye, but the eye with reported vision loss cannot reliably count fingers outside of 2 feet). Additional ocular symptoms associated with the vision loss, such as pain, redness, photophobia, new flashes or floaters, increase the urgency of the referral. The threshold for referral for any ocular complaint is lower compared with that of the general population for those with evidence of immunodeficiency, such as for this patient with HIV. Any CD4 count < 200 cells/mm³ should raise the practitioner’s concern for an ocular opportunistic infection, with the greatest concern with CD4 counts < 50 cells/mm³.

►Dr. Swamy. The patient underwent further testing in the ophthalmology clinic. Dr. Butler, can you please interpret the funduscopic exam?

►Dr. Butler. Both eyes demonstrate findings (microaneurysms and small dot-blot hemorrhages) consistent with moderate nonproliferative diabetic retinopathy (Figure 1A, white arrows). HIV-associated retinopathy could produce similar findings, but it is not generally seen with CD4 counts > 200 cells/mm³. Additionally, in the left eye, there is a diffuse patch of retinal whitening (retinitis) associated with the inferotemporal vascular arcades (Figure 1B, white arrows). The entire area involved is poorly circumscribed and the whitening is subtle in areas. Overlying some areas of deeper, ground-glass whitening there are scattered, punctate white spots (Figure 1B, green arrows). Wickremasinghe and colleagues described this pattern of retinitis and suggested that it had a high positive-predictive value in the diagnosis of ocular syphilis.⁵

►Dr. Swamy. The patient then underwent fluorescein angiography and optical coherence tomography (OCT). Dr. Butler, what did the fluorescein angiography show?

►Dr. Butler. The fluorescein angiogram in both eyes revealed leakage of dye consistent with diabetic retinopathy, with the right eye (OD) worse than the left (OS). Additionally, the areas of active retinitis in the left eye displayed gradual staining with leopard-spot changes, along with late leakage of fluorescein dye, indicating vasculopathy in the infected area (Figure 2, arrows). The patient also underwent OCT in the left eye (images not displayed) demonstrating vitreous cells (vitritis), patches of inner retinal thickening with hyperreflectivity, and hyperreflective nodules at the level of the retinal pigment epithelium with overlying photoreceptor disruption. These OCT findings are fairly stereotypic for syphilitic chorioretinitis.⁶

►Dr. Swamy. Based on the ophthalmic findings, a diagnosis of ocular syphilis was made. Dr. Serrao, what should internists consider as they evaluate and manage a patient with ocular syphilis?

►Dr. Serrao. Although isolated ocular involvement from syphilis is possible, the majority of patients (up to 85%) with HIV can present with concomitant central nervous system infection and about 30% present with symptomatic neurosyphilis (a typical late manifestation of this disease) that reflects the aggressiveness, accelerated course and propensity for wide dissemination of syphilis in this patient population.⁷

The presence of concomitant cutaneous rashes should prompt universal precautions, because transmission can occur via skin to skin contact. Clinicians should watch for the Jarisch-Herxheimer reaction during treatment, a syndrome of fever, myalgias, and headache, which results from circulating cytokines produced because of rapidly dying spirochetes that could mimic a penicillin drug reaction, yet is treated supportively.

As syphilis is sexually acquired, clinicians should test for coexistent sexually transmitted infections, vaccinate for those that are preventable (eg, hepatitis B), notify sexual partners via assistance from local departments of public health, and assess for coexistent drug use and offer counseling in order to optimize risk reduction. Special attention should be paid to virologic control of HIV since some studies have shown an increase in the propensity for breakthrough HIV viremia while on effective ART.⁹ This should warrant counseling for ongoing optimal ART adherence and close monitoring in the follow-up visits with a provider specialized in the treatment of syphilis and HIV.

►Dr. Swamy. A lumbar puncture is performed with the results listed in Table 2. Dr. Serrao, is the CSF consistent with neurosyphilis? What would you do next?

►Dr. Serrao. The lumbar puncture is inflammatory with a lymphocytic predominance, consistent with active ocular/neurosyphilis. The CSF Venereal Disease Research Laboratory test is specific but not sensitive so a negative value does not rule out the presence of central nervous system infection.¹⁰ The CSF fluorescent treponemal antibody (CSF FTA-ABS) is sensitive but not specific. In this case, the ocular findings, positive serum RPR, CSF lymphocytic predominance, and CSF FTA ABS strongly supports the diagnosis of ocular/early neurosyphilis in a patient with HIV infection in whom early aggressive treatment is warranted to prevent rapid progression/potential loss of vision.¹¹

►Dr. Swamy. Dr. Butler, how does syphilis behave in the eye as compared to other infectious or inflammatory diseases? Do visual symptoms respond well to treatment?

►Dr. Butler. As opposed to the dramatic reduction in rates and severity of CMV retinitis, HAART has had a negligible effect on ocular syphilis in the setting of HIV coinfection; in fact, rates of syphilis, including ocular syphilis, are currently surging world-wide, and HIV coinfection portends a worse prognosis.¹² This is especially true among gay men. More so, there appears to be no correlation between CD4 count and incidence of developing ocular syphilis, as opposed to CMV retinitis, which occurs far more frequently in those with CD4 counts < 50 cells/mm³. In keeping with its epithet as one of the “Great Imitators,” syphilis can affect virtually every tissue of the eye—conjunctiva, sclera, cornea, iris, lens, vitreous, retina, choroid, optic nerve—unlike other OOI, such as CMV or toxoplasma, which generally hone to the retina. Nonetheless, various findings and patterns on clinical exam and ancillary testing, such as the more recently described punctate inner retinitis (as seen in our patient) and the more classic acute syphilitic posterior placoid chorioretinitis, carry high specificity for ocular syphilis.¹³

Patients with ocular syphilis should be treated according to neurosyphilis treatment protocols. In general, these patients respond very well to treatment with resolution of the ocular findings and recovery of complete, or nearly so, visual function, as long as an excessive delay between diagnosis and proper treatment does not occur.¹⁴

►Dr. Swamy. Following this testing, the patient completed 14 days of IV penicillin with resolution of symptoms. He had no further vision complaints. He was started on Triumeq (abacavir, dolutegravir, and lamivudine) with good adherence to therapy. Dr. Serrao, in 2016 the CDC released a clinical advisory about ocular syphilis. Can you tell us about why this is an important diagnosis to be aware of today?

►Dr. Serrao. As with any disease, the epidemiologic characteristics of an infection like syphilis allow the clinician to more carefully entertain such a diagnosis in any one individual by improving the index of suspicion for a particular disease. Awareness of an increase in ocular syphilis in HIV positive MSM allows for a more timely assessment and subsequent treatment with the goal of preventing loss of vision.¹⁵

►Lakshmana Swamy, MD, chief medical resident, VA Boston Healthcare System (VABHS) and Boston Medical Center. Dr. Serrao, when you hear about vision changes in a patient with HIV, what differential diagnosis is generated? What epidemiologic or historical factors can help distinguish these entities?

►Richard Serrao, MD, Infectious Disease Service, VABHS and assistant professor of medicine, Boston University School of Medicine. The differential diagnoses for vision changes in a patient with HIV is based on the overall immunosuppression of the patient: the lower the patient’s CD4 count, the higher the number of etiologies.¹ The portions of the visual pathway as well as the pattern of vision loss are useful in narrowing the differential. For example, monocular visual disturbances with dermatomal vesicles within the ophthalmic division of the trigeminal nerve strongly implicates varicella zoster retinitis or keratitis; abducens nerve palsy could suggest granulomatous basilar meningitis from cryptococcosis. Likewise, ongoing fevers in an advanced AIDS patient with concomitant colitis, hepatitis, and pneumonitis is strongly suspicious for cytomegalovirus (CMV) retinitis with wide dissemination.

Geographic epidemiologic factors can suggest pathogens more prevalent to certain regions of the world, such as histoplasma chorioretinitis in a resident of the central and eastern U.S. or tuberculosis in a returning traveler. Likewise, a cat owner or one who consumes steak tartare increases the likelihood for toxoplasma retinochoroiditis, or syphilis in men who have sex with men (MSM) in the U.S. given that the majority of new cases occur in this patient population. Other clues one should consider include the presence of splinter hemorrhages in the extremities in an intravenous drug user, raising the possibility of embolic endophthalmitis from bacterial or fungal endocarditis. A variety of other diagnoses can certainly occur as a result of drug treatment (uveitis from rifampin, for example), immune reconstitution from HAART, infections with other HIV-associated pathogens, such as Pneumocystis jiroveci, and many non-HIV-related ocular diseases.

►Dr. Swamy. Dr. Butler, what concerns do you have when you hear about an HIV-infected patient with vision loss from the ophthalmology perspective?

►Nicholas Butler, MD, Ophthalmology Service, Uveitis and Ocular Immunology, VABHS and assistant professor of ophthalmology, Harvard Medical School. Of course, patients with HIV suffer from common causes of vision loss—cataract, glaucoma, diabetes, macular degeneration, for instance—just like those without HIV infection. If there is no significant immunodeficiency, then the patient’s HIV status would be less relevant, and these more common causes of vision loss should be pursued. My first task would be to determine the patient’s most recent CD4 T-cell count.

Assuming an HIV-positive individual is experiencing visual symptoms related to his/her underlying HIV infection (especially in the setting of CD4 counts < 200 cells/mm³), ocular opportunistic infections (OOI) come to mind first. Despite a reduction in incidence of 75% to 80% in the HAART-era, CMV retinitis remains the most common OOI in patients with AIDS and carries the greatest risk of ocular morbidity.² In fact, based on enrollment data for the Longitudinal Study of the Ocular Complications of AIDS (LSOCA), the prevalence of CMV retinitis among patients with AIDS is more than 20-fold higher than all other ocular complications of AIDS (OOIs and ocular neoplastic disease), including Kaposi sarcoma, lymphoma, herpes zoster ophthalmicus, ocular syphilis, ocular toxoplasma, necrotizing herpetic retinitis, cryptococcal choroiditis, and pneumocystis choroiditis.³ Beyond ocular opportunistic infections, the most common retinal finding in HIV-positive people is HIV retinopathy, nonspecific microvascular findings in the retina affecting nearly 70% of those with advanced HIV disease. Fortunately, HIV retinopathy is generally asymptomatic.⁴

►Dr. Swamy. Thank you for those explanations. Based on Dr. Serrao’s differential, it is worth noting that this patient is MSM. He was evaluated in urgent care with the initial examination showing a temperature of 98.0° F, pulse 83 beats per minute, and blood pressure 110/70 mm Hg. The eye exam showed no injection with normal extraocular movements. Initial laboratory data were notable for a CD4 count of 730 cells/mm³ with fewer than 20 HIV viral copies/mL. Cytomegalovirus immunoglobulin G (IgG) was positive, and immunoglobulin M (IgM) was negative. A Lyme antibody was positive with negative IgM and IgG by Western blot. Additional tests can be seen in Tables 1 and 2. The patient has good immunologic and virologic control. How does this change your thinking about the case?

►Dr. Serrao. His CD4 count is well above 350, increasing the likelihood of a relatively uncomplicated course and treatment. Cytomegalovirus antibodies reflect prior infection. As CMV generally does not manifest with disease of any variety (including CMV retinitis) at this high CD4 count, one can presume he does not have CMV retinitis as a cause for his visual changes. CMV retinitis occurs mainly when substantial CD4 depletion has occurred (typically less than 50 cells/mm³). A positive Lyme antibody screen, not specific to Lyme, can be falsely positive in other treponema diseases (eg, Treponema pallidum, the etiologic organism of syphilis) as evidenced by negative confirmatory Western blot IgG and IgM. Antineutrophil cystoplasmic antibodies, lysozyme, angiotensin-converting enzyme, rapid plasma reagin (RPR), herpes simplex virus, toxoplasma are generally included in the workup for the differential of uveitis, retinitis, choroiditis, etc.

►Dr. Swamy. Based on the visual changes, the patient was referred for urgent ophthalmologic evaluation. Dr. Butler, when should a generalist consider urgent ophthalmology referral?

►Dr. Butler. In general, all patients with acute (and significant) vision loss should be referred immediately to an ophthalmologist. The challenge for the general practitioner is determining the true extent of the reported vision loss. If possible, some assessment of visual acuity should be obtained, testing each eye independently and with the correct glasses correction (ie, the patient’s distance glasses if the test object is 12 feet or more from the patient or their reading glasses if the test object is held inside arm’s length). If the general practitioner does not have access to an eye chart or near card, any assessment of vision with an appropriate description will be useful (eg, the patient can quickly count fingers at 15 feet in the unaffected eye, but the eye with reported vision loss cannot reliably count fingers outside of 2 feet). Additional ocular symptoms associated with the vision loss, such as pain, redness, photophobia, new flashes or floaters, increase the urgency of the referral. The threshold for referral for any ocular complaint is lower compared with that of the general population for those with evidence of immunodeficiency, such as for this patient with HIV. Any CD4 count < 200 cells/mm³ should raise the practitioner’s concern for an ocular opportunistic infection, with the greatest concern with CD4 counts < 50 cells/mm³.

►Dr. Swamy. The patient underwent further testing in the ophthalmology clinic. Dr. Butler, can you please interpret the funduscopic exam?

►Dr. Butler. Both eyes demonstrate findings (microaneurysms and small dot-blot hemorrhages) consistent with moderate nonproliferative diabetic retinopathy (Figure 1A, white arrows). HIV-associated retinopathy could produce similar findings, but it is not generally seen with CD4 counts > 200 cells/mm³. Additionally, in the left eye, there is a diffuse patch of retinal whitening (retinitis) associated with the inferotemporal vascular arcades (Figure 1B, white arrows). The entire area involved is poorly circumscribed and the whitening is subtle in areas. Overlying some areas of deeper, ground-glass whitening there are scattered, punctate white spots (Figure 1B, green arrows). Wickremasinghe and colleagues described this pattern of retinitis and suggested that it had a high positive-predictive value in the diagnosis of ocular syphilis.⁵

►Dr. Swamy. The patient then underwent fluorescein angiography and optical coherence tomography (OCT). Dr. Butler, what did the fluorescein angiography show?

►Dr. Butler. The fluorescein angiogram in both eyes revealed leakage of dye consistent with diabetic retinopathy, with the right eye (OD) worse than the left (OS). Additionally, the areas of active retinitis in the left eye displayed gradual staining with leopard-spot changes, along with late leakage of fluorescein dye, indicating vasculopathy in the infected area (Figure 2, arrows). The patient also underwent OCT in the left eye (images not displayed) demonstrating vitreous cells (vitritis), patches of inner retinal thickening with hyperreflectivity, and hyperreflective nodules at the level of the retinal pigment epithelium with overlying photoreceptor disruption. These OCT findings are fairly stereotypic for syphilitic chorioretinitis.⁶

►Dr. Swamy. Based on the ophthalmic findings, a diagnosis of ocular syphilis was made. Dr. Serrao, what should internists consider as they evaluate and manage a patient with ocular syphilis?

►Dr. Serrao. Although isolated ocular involvement from syphilis is possible, the majority of patients (up to 85%) with HIV can present with concomitant central nervous system infection and about 30% present with symptomatic neurosyphilis (a typical late manifestation of this disease) that reflects the aggressiveness, accelerated course and propensity for wide dissemination of syphilis in this patient population.⁷

The presence of concomitant cutaneous rashes should prompt universal precautions, because transmission can occur via skin to skin contact. Clinicians should watch for the Jarisch-Herxheimer reaction during treatment, a syndrome of fever, myalgias, and headache, which results from circulating cytokines produced because of rapidly dying spirochetes that could mimic a penicillin drug reaction, yet is treated supportively.

As syphilis is sexually acquired, clinicians should test for coexistent sexually transmitted infections, vaccinate for those that are preventable (eg, hepatitis B), notify sexual partners via assistance from local departments of public health, and assess for coexistent drug use and offer counseling in order to optimize risk reduction. Special attention should be paid to virologic control of HIV since some studies have shown an increase in the propensity for breakthrough HIV viremia while on effective ART.⁹ This should warrant counseling for ongoing optimal ART adherence and close monitoring in the follow-up visits with a provider specialized in the treatment of syphilis and HIV.

►Dr. Swamy. A lumbar puncture is performed with the results listed in Table 2. Dr. Serrao, is the CSF consistent with neurosyphilis? What would you do next?

►Dr. Serrao. The lumbar puncture is inflammatory with a lymphocytic predominance, consistent with active ocular/neurosyphilis. The CSF Venereal Disease Research Laboratory test is specific but not sensitive so a negative value does not rule out the presence of central nervous system infection.¹⁰ The CSF fluorescent treponemal antibody (CSF FTA-ABS) is sensitive but not specific. In this case, the ocular findings, positive serum RPR, CSF lymphocytic predominance, and CSF FTA ABS strongly supports the diagnosis of ocular/early neurosyphilis in a patient with HIV infection in whom early aggressive treatment is warranted to prevent rapid progression/potential loss of vision.¹¹

►Dr. Swamy. Dr. Butler, how does syphilis behave in the eye as compared to other infectious or inflammatory diseases? Do visual symptoms respond well to treatment?

►Dr. Butler. As opposed to the dramatic reduction in rates and severity of CMV retinitis, HAART has had a negligible effect on ocular syphilis in the setting of HIV coinfection; in fact, rates of syphilis, including ocular syphilis, are currently surging world-wide, and HIV coinfection portends a worse prognosis.¹² This is especially true among gay men. More so, there appears to be no correlation between CD4 count and incidence of developing ocular syphilis, as opposed to CMV retinitis, which occurs far more frequently in those with CD4 counts < 50 cells/mm³. In keeping with its epithet as one of the “Great Imitators,” syphilis can affect virtually every tissue of the eye—conjunctiva, sclera, cornea, iris, lens, vitreous, retina, choroid, optic nerve—unlike other OOI, such as CMV or toxoplasma, which generally hone to the retina. Nonetheless, various findings and patterns on clinical exam and ancillary testing, such as the more recently described punctate inner retinitis (as seen in our patient) and the more classic acute syphilitic posterior placoid chorioretinitis, carry high specificity for ocular syphilis.¹³

Patients with ocular syphilis should be treated according to neurosyphilis treatment protocols. In general, these patients respond very well to treatment with resolution of the ocular findings and recovery of complete, or nearly so, visual function, as long as an excessive delay between diagnosis and proper treatment does not occur.¹⁴

►Dr. Swamy. Following this testing, the patient completed 14 days of IV penicillin with resolution of symptoms. He had no further vision complaints. He was started on Triumeq (abacavir, dolutegravir, and lamivudine) with good adherence to therapy. Dr. Serrao, in 2016 the CDC released a clinical advisory about ocular syphilis. Can you tell us about why this is an important diagnosis to be aware of today?

►Dr. Serrao. As with any disease, the epidemiologic characteristics of an infection like syphilis allow the clinician to more carefully entertain such a diagnosis in any one individual by improving the index of suspicion for a particular disease. Awareness of an increase in ocular syphilis in HIV positive MSM allows for a more timely assessment and subsequent treatment with the goal of preventing loss of vision.¹⁵

Use of methamphetamine, an N-methyl analog of amphetamine, is a serious public health problem; throughout the world an estimated 35.7 million people use the drug recreationally.¹ Methamphetamine is easy to obtain because it is cheap to produce and can be synthesized anywhere. In the United States, methamphetamine is commonly manufactured in small-scale laboratories using relatively inexpensive, legally available ingredients. Large-scale manufacturing in clandestine laboratories also contributes to methamphetamine abuse. The drug, known as meth, crystal meth, ice, and other names, is available as a powder, tablet, or crystalline salt, and is used by various routes of administration (Table).
 

• provide a brief synopsis of the clinical presentation of patients who use methamphetamine
• describe some of the complications of methamphetamine abuse/dependence, focusing on methamphetamine-induced psychosis
• suggest ways to approach the treatment of these patients, including those with methamphetamine-induced psychosis.

Acute effects of methamphetamine use

Psychiatric symptoms. Patients under the influence of methamphetamine may present with clinical symptoms that mimic psychiatric disorders. For example, the drug can cause marked euphoria, hyperactivity, and disturbed speech patterns, thus mimicking a manic state. Patients also may present with anxiety, agitation, and irritability or aggressiveness. Although an individual may take methamphetamine for sexual enhancement, the drug can cause hyper­sexuality, which often is associated with unintended and unsafe sexual activities. These signs and symptoms are exacerbated during drug binges that can last for days, during which time large quantities of the drug are consumed.

Methamphetamine users may become preoccupied with their own thought patterns, and their actions can become compulsive and nonsensical. For example, a patient may become obsessed with an object of no specific value in his (her) environment, such as a doorknob or a cloud. Patients also may become suspicious of their friends and family members or think that police officers are after them. Less commonly, a patient also may suffer from poverty of speech, psychomotor retardation, and diminished social engagement similar to that reported in some patients with schizophrenia with deficit syndrome. Usually, acute symptoms will last 4 to 7 days after drug cessation, and then resolve completely with protracted abstinence from the drug.

Neurologic signs of methamphetamine use include hemorrhagic strokes in young people without any evidence of previous neurologic impairments. Studies have documented similarities between methamphetamine-induced neurotoxicity and traumatic brain injury.⁵ Postmortem studies have reported the presence of arteriovenous malformation in some patients with hemorrhagic strokes.

Hyperthermia is a dangerous acute effect of methamphetamine use. High body temperatures can cause both peripheral and central abnormalities, including muscular and cardiovascular dysfunction, renal failure secondary to rhabdomyolysis, heat stroke, and other heat-induced malignant syndromes. Some of the central dysfunctions may be related to heat-induced production of free radicals in various brain regions. There are no pharmacologic treatments for methamphetamine-induced thermal dysregulation.⁶ Therefore, clinicians need to focus on reducing body temperature by using cooling fans or cold water baths. Efforts should be made to avoid overhydrating patients because of the risk of developing the syndrome of inappropriate antidiuretic hormone secretion.

Chronic methamphetamine abuse

Psychosis is a long-term complication of chronic abuse of the drug.⁷ Although psychosis has been a reported complication of methamphetamine use since the 1950s,⁸ most of the subsequent literature is from Japan, where methamphetamine use was highly prevalent after World War II.^9,10 The prevalence of methamphetamine-induced psychosis in methamphetamine-dependent patients varies from 13% (in the United States¹¹) to 50% (in Asia¹²). This difference might be related to variability in the purity of methamphetamine used in different locations.

Methamphetamine users may experience a pre-psychotic state that consists of ideas of reference and delusional moods. This is followed by a psychotic state that includes hallucinations and delusions. The time it takes to develop these symptoms can vary from a few months up to >20 years after starting to use methamphetamine.^10,13 Psychosis can occur in patients who do not have a history of psychiatric illness.¹⁰

The clinical presentation of methamphetamine-induced psychosis includes delusions of reference and persecutions.^8-10 Paranoid delusions may be accompanied by violent behavior. Some patients may present with grandiose or jealousy delusions. Patients may experience auditory, tactile, or visual hallucinations. They may exhibit mania and logorrheic verbal outputs, symptoms consistent with a diagnosis of methamphetamine-induced mood disorder with manic features. Patients who use large daily doses of the drug also may report that there are ants or other parasites crawling under their skin (eg, formication, “meth mites”) and might present with infected excoriations of their skin as a result of attempting to remove insects. This is clinically important because penicillin-resistant bacteria are common in patients who use methamphetamine, and strains tend to be virulent.

Psychotic symptoms can last from a few days to several weeks after stopping methamphetamine use, although methamphetamine-induced psychosis can persist after long periods of abstinence.¹⁴ Psychotic symptoms may recur with re-exposure to the drug⁹ or repeated stressful life events.¹⁵ Patients with recurrent psychosis in the absence of a drug trigger appear to have high levels of peripheral norepinephrine.¹⁵ Patients with psychosis caused by long-term methamphetamine use will not necessarily show signs of sympathomimetic dysfunction because they may not have any methamphetamine in the body when they first present for clinical evaluation. Importantly, patients with methamphetamine-induced psychosis have been reported to have poor outcomes at follow-up.¹⁶ They have an increased risk of suicide, recurrent drug-induced psychosis, and comorbid alcohol abuse.¹⁶

Doses required to induce psychosis vary from patient to patient and may depend on the patient’s genetic background and/or environmental conditions. Methamphetamine can increase the severity of many psychiatric symptoms¹⁷ and may expedite the development of schizophrenia in first-degree relatives of patients with schizophrenia.¹⁸

The diagnosis of methamphetamine-induced psychosis should focus on differentiating it from schizophrenia. Wang et al¹⁹ found similar patterns of delusions in patients with schizophrenia and those with methamphetamine-induced psychosis. However, compared with patients with schizophrenia, patients with methamphetamine-induced psychosis have a higher prevalence of visual and tactile hallucinations, and less disorganization, blunted affect, and motor retardation. Some patients may present with depression and suicidal ideation; these features may be more prominent during withdrawal, but also may be obvious during periods of active use.¹⁶

Although these clinical features may be helpful initially, more comparative neurobiologic investigations are needed to identify potential biologic differences between schizophrenia and methamphetamine-induced psychosis because these differences will impact therapeutic approaches to these diverse population groups.

Neurologic complications. Chronic methamphetamine users may develop various neurologic disorders.²⁰ They may present with stereotypies involving finger movements or repeated rubbing of mouth or face, orofacial dyskinesia, and choreoathetoid movements reminiscent of classical neurologic disorders. These movement disorders can persist after cessation of methamphetamine use. In some cases, these movement abnormalities may respond to dopamine receptor antagonists such as haloperidol.

Neuropsychological findings. Chronic methamphetamine users show mild signs of cognitive decline that affects a broad range of neuropsychological functions.^21-23 There are deficits in several cognitive processes that are dependent on the function of frontostriatal and limbic circuits.^24-26 Specifically, episodic memory, executive functions, complex information processing speed, and psychomotor functions all have been reported to be negatively impacted.

Methamphetamine use often results in psychiatric distress that impacts users’ interpersonal relationships.²⁷ Additionally, impulsivity may exacerbate their psychosocial difficulties and promote maintenance of drug-seeking behaviors.²⁸ Cognitive deficits lead to poor health outcomes, high-risk behaviors, employment difficulties, and repeated relapse.^29,30

Partial recovery of neuropsychological functioning and improvement in affective distress can be achieved after sustained abstinence from methamphetamine, but recovery may not be complete. Because cognitive dysfunction can influence treatment outcomes, clinicians need to be fully aware of the cognitive status of those patients, and a thorough neuropsychological evaluation is necessary before initiating treatment.

Treatment

Methamphetamine abuse. Because patients who abuse methamphetamine are at high risk of developing psychosis, neuro­logic complications, and neuropsychological disorders, initiating treatment early in the course of their addiction is of paramount importance. Treatment of methamphetamine addiction is complicated by the fact that these patients have a high prevalence of comorbid psychiatric disorders, which clinicians need to keep in mind when selecting therapeutic interventions.

There are no FDA-approved agents for treating methamphetamine abuse.³¹ Several drugs have been tried with varying degrees of success, including bupropion, modafinil, and naltrexone. A study of modafinil found no clinically significant effects for treating methamphetamine abuse; however, only approximately one-half of participants in this study took modafinil as instructed.³² Certain selective serotonin reuptake inhibitors, including fluoxetine and paroxetine, have not been shown to be effective in treating these patients. Naltrexone may be a reasonable medication to consider because of the high prevalence of comorbid alcohol abuse among methamphetamine users.

Other treatments for methamphetamine addiction consist of behavioral interventions such as cognitive-behavioral therapy. Clinical experience has shown that the risk of relapse depends on how long the patient has been abstinent prior to entering a treatment program, the presence of attention and memory deficits, and findings of poor decision-making on neuropsychological tests.

The presence of cognitive abnormalities has been reported to impact methamphetamine abusers’ response to treatment.³³ These findings suggest the need to develop approaches that might improve cognition in patients who are undergoing treatment for methamphetamine abuse. The monoaminergic agent modafinil and similar drugs need to be evaluated in large populations to increase the possibility of identifying characteristics of patients who might respond to cognitive enhancement.³⁴

Methamphetamine-induced psychosis. First-generation antipsychotics, such as haloperidol or fluphenazine, need to be used sparingly in patients with methamphetamine-induced psychosis because of the risk of developing extrapyramidal symptoms (EPS) and because these patients are prone to develop motor complications as a result of methamphetamine abuse. Second-generation antipsychotics, such as risperidone and olanzapine, may be more appropriate because of the lower risks of EPS.³⁵ The presence of high norepinephrine levels in some patients with recurrent methamphetamine psychosis suggests that drugs that block norepinephrine receptors, such as prazosin or propranolol, might be of therapeutic benefit if they are shown to be effective in controlled clinical trials.

Use of methamphetamine, an N-methyl analog of amphetamine, is a serious public health problem; throughout the world an estimated 35.7 million people use the drug recreationally.¹ Methamphetamine is easy to obtain because it is cheap to produce and can be synthesized anywhere. In the United States, methamphetamine is commonly manufactured in small-scale laboratories using relatively inexpensive, legally available ingredients. Large-scale manufacturing in clandestine laboratories also contributes to methamphetamine abuse. The drug, known as meth, crystal meth, ice, and other names, is available as a powder, tablet, or crystalline salt, and is used by various routes of administration (Table).
 

• provide a brief synopsis of the clinical presentation of patients who use methamphetamine
• describe some of the complications of methamphetamine abuse/dependence, focusing on methamphetamine-induced psychosis
• suggest ways to approach the treatment of these patients, including those with methamphetamine-induced psychosis.

Acute effects of methamphetamine use

Psychiatric symptoms. Patients under the influence of methamphetamine may present with clinical symptoms that mimic psychiatric disorders. For example, the drug can cause marked euphoria, hyperactivity, and disturbed speech patterns, thus mimicking a manic state. Patients also may present with anxiety, agitation, and irritability or aggressiveness. Although an individual may take methamphetamine for sexual enhancement, the drug can cause hyper­sexuality, which often is associated with unintended and unsafe sexual activities. These signs and symptoms are exacerbated during drug binges that can last for days, during which time large quantities of the drug are consumed.

Methamphetamine users may become preoccupied with their own thought patterns, and their actions can become compulsive and nonsensical. For example, a patient may become obsessed with an object of no specific value in his (her) environment, such as a doorknob or a cloud. Patients also may become suspicious of their friends and family members or think that police officers are after them. Less commonly, a patient also may suffer from poverty of speech, psychomotor retardation, and diminished social engagement similar to that reported in some patients with schizophrenia with deficit syndrome. Usually, acute symptoms will last 4 to 7 days after drug cessation, and then resolve completely with protracted abstinence from the drug.

Neurologic signs of methamphetamine use include hemorrhagic strokes in young people without any evidence of previous neurologic impairments. Studies have documented similarities between methamphetamine-induced neurotoxicity and traumatic brain injury.⁵ Postmortem studies have reported the presence of arteriovenous malformation in some patients with hemorrhagic strokes.

Hyperthermia is a dangerous acute effect of methamphetamine use. High body temperatures can cause both peripheral and central abnormalities, including muscular and cardiovascular dysfunction, renal failure secondary to rhabdomyolysis, heat stroke, and other heat-induced malignant syndromes. Some of the central dysfunctions may be related to heat-induced production of free radicals in various brain regions. There are no pharmacologic treatments for methamphetamine-induced thermal dysregulation.⁶ Therefore, clinicians need to focus on reducing body temperature by using cooling fans or cold water baths. Efforts should be made to avoid overhydrating patients because of the risk of developing the syndrome of inappropriate antidiuretic hormone secretion.

Chronic methamphetamine abuse

Psychosis is a long-term complication of chronic abuse of the drug.⁷ Although psychosis has been a reported complication of methamphetamine use since the 1950s,⁸ most of the subsequent literature is from Japan, where methamphetamine use was highly prevalent after World War II.^9,10 The prevalence of methamphetamine-induced psychosis in methamphetamine-dependent patients varies from 13% (in the United States¹¹) to 50% (in Asia¹²). This difference might be related to variability in the purity of methamphetamine used in different locations.

Methamphetamine users may experience a pre-psychotic state that consists of ideas of reference and delusional moods. This is followed by a psychotic state that includes hallucinations and delusions. The time it takes to develop these symptoms can vary from a few months up to >20 years after starting to use methamphetamine.^10,13 Psychosis can occur in patients who do not have a history of psychiatric illness.¹⁰

The clinical presentation of methamphetamine-induced psychosis includes delusions of reference and persecutions.^8-10 Paranoid delusions may be accompanied by violent behavior. Some patients may present with grandiose or jealousy delusions. Patients may experience auditory, tactile, or visual hallucinations. They may exhibit mania and logorrheic verbal outputs, symptoms consistent with a diagnosis of methamphetamine-induced mood disorder with manic features. Patients who use large daily doses of the drug also may report that there are ants or other parasites crawling under their skin (eg, formication, “meth mites”) and might present with infected excoriations of their skin as a result of attempting to remove insects. This is clinically important because penicillin-resistant bacteria are common in patients who use methamphetamine, and strains tend to be virulent.

Psychotic symptoms can last from a few days to several weeks after stopping methamphetamine use, although methamphetamine-induced psychosis can persist after long periods of abstinence.¹⁴ Psychotic symptoms may recur with re-exposure to the drug⁹ or repeated stressful life events.¹⁵ Patients with recurrent psychosis in the absence of a drug trigger appear to have high levels of peripheral norepinephrine.¹⁵ Patients with psychosis caused by long-term methamphetamine use will not necessarily show signs of sympathomimetic dysfunction because they may not have any methamphetamine in the body when they first present for clinical evaluation. Importantly, patients with methamphetamine-induced psychosis have been reported to have poor outcomes at follow-up.¹⁶ They have an increased risk of suicide, recurrent drug-induced psychosis, and comorbid alcohol abuse.¹⁶

Doses required to induce psychosis vary from patient to patient and may depend on the patient’s genetic background and/or environmental conditions. Methamphetamine can increase the severity of many psychiatric symptoms¹⁷ and may expedite the development of schizophrenia in first-degree relatives of patients with schizophrenia.¹⁸

The diagnosis of methamphetamine-induced psychosis should focus on differentiating it from schizophrenia. Wang et al¹⁹ found similar patterns of delusions in patients with schizophrenia and those with methamphetamine-induced psychosis. However, compared with patients with schizophrenia, patients with methamphetamine-induced psychosis have a higher prevalence of visual and tactile hallucinations, and less disorganization, blunted affect, and motor retardation. Some patients may present with depression and suicidal ideation; these features may be more prominent during withdrawal, but also may be obvious during periods of active use.¹⁶

Although these clinical features may be helpful initially, more comparative neurobiologic investigations are needed to identify potential biologic differences between schizophrenia and methamphetamine-induced psychosis because these differences will impact therapeutic approaches to these diverse population groups.

Neurologic complications. Chronic methamphetamine users may develop various neurologic disorders.²⁰ They may present with stereotypies involving finger movements or repeated rubbing of mouth or face, orofacial dyskinesia, and choreoathetoid movements reminiscent of classical neurologic disorders. These movement disorders can persist after cessation of methamphetamine use. In some cases, these movement abnormalities may respond to dopamine receptor antagonists such as haloperidol.

Neuropsychological findings. Chronic methamphetamine users show mild signs of cognitive decline that affects a broad range of neuropsychological functions.^21-23 There are deficits in several cognitive processes that are dependent on the function of frontostriatal and limbic circuits.^24-26 Specifically, episodic memory, executive functions, complex information processing speed, and psychomotor functions all have been reported to be negatively impacted.

Methamphetamine use often results in psychiatric distress that impacts users’ interpersonal relationships.²⁷ Additionally, impulsivity may exacerbate their psychosocial difficulties and promote maintenance of drug-seeking behaviors.²⁸ Cognitive deficits lead to poor health outcomes, high-risk behaviors, employment difficulties, and repeated relapse.^29,30

Partial recovery of neuropsychological functioning and improvement in affective distress can be achieved after sustained abstinence from methamphetamine, but recovery may not be complete. Because cognitive dysfunction can influence treatment outcomes, clinicians need to be fully aware of the cognitive status of those patients, and a thorough neuropsychological evaluation is necessary before initiating treatment.

Treatment

Methamphetamine abuse. Because patients who abuse methamphetamine are at high risk of developing psychosis, neuro­logic complications, and neuropsychological disorders, initiating treatment early in the course of their addiction is of paramount importance. Treatment of methamphetamine addiction is complicated by the fact that these patients have a high prevalence of comorbid psychiatric disorders, which clinicians need to keep in mind when selecting therapeutic interventions.

There are no FDA-approved agents for treating methamphetamine abuse.³¹ Several drugs have been tried with varying degrees of success, including bupropion, modafinil, and naltrexone. A study of modafinil found no clinically significant effects for treating methamphetamine abuse; however, only approximately one-half of participants in this study took modafinil as instructed.³² Certain selective serotonin reuptake inhibitors, including fluoxetine and paroxetine, have not been shown to be effective in treating these patients. Naltrexone may be a reasonable medication to consider because of the high prevalence of comorbid alcohol abuse among methamphetamine users.

Other treatments for methamphetamine addiction consist of behavioral interventions such as cognitive-behavioral therapy. Clinical experience has shown that the risk of relapse depends on how long the patient has been abstinent prior to entering a treatment program, the presence of attention and memory deficits, and findings of poor decision-making on neuropsychological tests.

The presence of cognitive abnormalities has been reported to impact methamphetamine abusers’ response to treatment.³³ These findings suggest the need to develop approaches that might improve cognition in patients who are undergoing treatment for methamphetamine abuse. The monoaminergic agent modafinil and similar drugs need to be evaluated in large populations to increase the possibility of identifying characteristics of patients who might respond to cognitive enhancement.³⁴

Methamphetamine-induced psychosis. First-generation antipsychotics, such as haloperidol or fluphenazine, need to be used sparingly in patients with methamphetamine-induced psychosis because of the risk of developing extrapyramidal symptoms (EPS) and because these patients are prone to develop motor complications as a result of methamphetamine abuse. Second-generation antipsychotics, such as risperidone and olanzapine, may be more appropriate because of the lower risks of EPS.³⁵ The presence of high norepinephrine levels in some patients with recurrent methamphetamine psychosis suggests that drugs that block norepinephrine receptors, such as prazosin or propranolol, might be of therapeutic benefit if they are shown to be effective in controlled clinical trials.

Use of methamphetamine, an N-methyl analog of amphetamine, is a serious public health problem; throughout the world an estimated 35.7 million people use the drug recreationally.¹ Methamphetamine is easy to obtain because it is cheap to produce and can be synthesized anywhere. In the United States, methamphetamine is commonly manufactured in small-scale laboratories using relatively inexpensive, legally available ingredients. Large-scale manufacturing in clandestine laboratories also contributes to methamphetamine abuse. The drug, known as meth, crystal meth, ice, and other names, is available as a powder, tablet, or crystalline salt, and is used by various routes of administration (Table).
 

• provide a brief synopsis of the clinical presentation of patients who use methamphetamine
• describe some of the complications of methamphetamine abuse/dependence, focusing on methamphetamine-induced psychosis
• suggest ways to approach the treatment of these patients, including those with methamphetamine-induced psychosis.

Acute effects of methamphetamine use

Psychiatric symptoms. Patients under the influence of methamphetamine may present with clinical symptoms that mimic psychiatric disorders. For example, the drug can cause marked euphoria, hyperactivity, and disturbed speech patterns, thus mimicking a manic state. Patients also may present with anxiety, agitation, and irritability or aggressiveness. Although an individual may take methamphetamine for sexual enhancement, the drug can cause hyper­sexuality, which often is associated with unintended and unsafe sexual activities. These signs and symptoms are exacerbated during drug binges that can last for days, during which time large quantities of the drug are consumed.

Methamphetamine users may become preoccupied with their own thought patterns, and their actions can become compulsive and nonsensical. For example, a patient may become obsessed with an object of no specific value in his (her) environment, such as a doorknob or a cloud. Patients also may become suspicious of their friends and family members or think that police officers are after them. Less commonly, a patient also may suffer from poverty of speech, psychomotor retardation, and diminished social engagement similar to that reported in some patients with schizophrenia with deficit syndrome. Usually, acute symptoms will last 4 to 7 days after drug cessation, and then resolve completely with protracted abstinence from the drug.

Neurologic signs of methamphetamine use include hemorrhagic strokes in young people without any evidence of previous neurologic impairments. Studies have documented similarities between methamphetamine-induced neurotoxicity and traumatic brain injury.⁵ Postmortem studies have reported the presence of arteriovenous malformation in some patients with hemorrhagic strokes.

Hyperthermia is a dangerous acute effect of methamphetamine use. High body temperatures can cause both peripheral and central abnormalities, including muscular and cardiovascular dysfunction, renal failure secondary to rhabdomyolysis, heat stroke, and other heat-induced malignant syndromes. Some of the central dysfunctions may be related to heat-induced production of free radicals in various brain regions. There are no pharmacologic treatments for methamphetamine-induced thermal dysregulation.⁶ Therefore, clinicians need to focus on reducing body temperature by using cooling fans or cold water baths. Efforts should be made to avoid overhydrating patients because of the risk of developing the syndrome of inappropriate antidiuretic hormone secretion.

Chronic methamphetamine abuse

Psychosis is a long-term complication of chronic abuse of the drug.⁷ Although psychosis has been a reported complication of methamphetamine use since the 1950s,⁸ most of the subsequent literature is from Japan, where methamphetamine use was highly prevalent after World War II.^9,10 The prevalence of methamphetamine-induced psychosis in methamphetamine-dependent patients varies from 13% (in the United States¹¹) to 50% (in Asia¹²). This difference might be related to variability in the purity of methamphetamine used in different locations.

Methamphetamine users may experience a pre-psychotic state that consists of ideas of reference and delusional moods. This is followed by a psychotic state that includes hallucinations and delusions. The time it takes to develop these symptoms can vary from a few months up to >20 years after starting to use methamphetamine.^10,13 Psychosis can occur in patients who do not have a history of psychiatric illness.¹⁰

The clinical presentation of methamphetamine-induced psychosis includes delusions of reference and persecutions.^8-10 Paranoid delusions may be accompanied by violent behavior. Some patients may present with grandiose or jealousy delusions. Patients may experience auditory, tactile, or visual hallucinations. They may exhibit mania and logorrheic verbal outputs, symptoms consistent with a diagnosis of methamphetamine-induced mood disorder with manic features. Patients who use large daily doses of the drug also may report that there are ants or other parasites crawling under their skin (eg, formication, “meth mites”) and might present with infected excoriations of their skin as a result of attempting to remove insects. This is clinically important because penicillin-resistant bacteria are common in patients who use methamphetamine, and strains tend to be virulent.

Psychotic symptoms can last from a few days to several weeks after stopping methamphetamine use, although methamphetamine-induced psychosis can persist after long periods of abstinence.¹⁴ Psychotic symptoms may recur with re-exposure to the drug⁹ or repeated stressful life events.¹⁵ Patients with recurrent psychosis in the absence of a drug trigger appear to have high levels of peripheral norepinephrine.¹⁵ Patients with psychosis caused by long-term methamphetamine use will not necessarily show signs of sympathomimetic dysfunction because they may not have any methamphetamine in the body when they first present for clinical evaluation. Importantly, patients with methamphetamine-induced psychosis have been reported to have poor outcomes at follow-up.¹⁶ They have an increased risk of suicide, recurrent drug-induced psychosis, and comorbid alcohol abuse.¹⁶

Doses required to induce psychosis vary from patient to patient and may depend on the patient’s genetic background and/or environmental conditions. Methamphetamine can increase the severity of many psychiatric symptoms¹⁷ and may expedite the development of schizophrenia in first-degree relatives of patients with schizophrenia.¹⁸

The diagnosis of methamphetamine-induced psychosis should focus on differentiating it from schizophrenia. Wang et al¹⁹ found similar patterns of delusions in patients with schizophrenia and those with methamphetamine-induced psychosis. However, compared with patients with schizophrenia, patients with methamphetamine-induced psychosis have a higher prevalence of visual and tactile hallucinations, and less disorganization, blunted affect, and motor retardation. Some patients may present with depression and suicidal ideation; these features may be more prominent during withdrawal, but also may be obvious during periods of active use.¹⁶

Although these clinical features may be helpful initially, more comparative neurobiologic investigations are needed to identify potential biologic differences between schizophrenia and methamphetamine-induced psychosis because these differences will impact therapeutic approaches to these diverse population groups.

Neurologic complications. Chronic methamphetamine users may develop various neurologic disorders.²⁰ They may present with stereotypies involving finger movements or repeated rubbing of mouth or face, orofacial dyskinesia, and choreoathetoid movements reminiscent of classical neurologic disorders. These movement disorders can persist after cessation of methamphetamine use. In some cases, these movement abnormalities may respond to dopamine receptor antagonists such as haloperidol.

Neuropsychological findings. Chronic methamphetamine users show mild signs of cognitive decline that affects a broad range of neuropsychological functions.^21-23 There are deficits in several cognitive processes that are dependent on the function of frontostriatal and limbic circuits.^24-26 Specifically, episodic memory, executive functions, complex information processing speed, and psychomotor functions all have been reported to be negatively impacted.

Methamphetamine use often results in psychiatric distress that impacts users’ interpersonal relationships.²⁷ Additionally, impulsivity may exacerbate their psychosocial difficulties and promote maintenance of drug-seeking behaviors.²⁸ Cognitive deficits lead to poor health outcomes, high-risk behaviors, employment difficulties, and repeated relapse.^29,30

Partial recovery of neuropsychological functioning and improvement in affective distress can be achieved after sustained abstinence from methamphetamine, but recovery may not be complete. Because cognitive dysfunction can influence treatment outcomes, clinicians need to be fully aware of the cognitive status of those patients, and a thorough neuropsychological evaluation is necessary before initiating treatment.

Treatment

Methamphetamine abuse. Because patients who abuse methamphetamine are at high risk of developing psychosis, neuro­logic complications, and neuropsychological disorders, initiating treatment early in the course of their addiction is of paramount importance. Treatment of methamphetamine addiction is complicated by the fact that these patients have a high prevalence of comorbid psychiatric disorders, which clinicians need to keep in mind when selecting therapeutic interventions.

There are no FDA-approved agents for treating methamphetamine abuse.³¹ Several drugs have been tried with varying degrees of success, including bupropion, modafinil, and naltrexone. A study of modafinil found no clinically significant effects for treating methamphetamine abuse; however, only approximately one-half of participants in this study took modafinil as instructed.³² Certain selective serotonin reuptake inhibitors, including fluoxetine and paroxetine, have not been shown to be effective in treating these patients. Naltrexone may be a reasonable medication to consider because of the high prevalence of comorbid alcohol abuse among methamphetamine users.

Other treatments for methamphetamine addiction consist of behavioral interventions such as cognitive-behavioral therapy. Clinical experience has shown that the risk of relapse depends on how long the patient has been abstinent prior to entering a treatment program, the presence of attention and memory deficits, and findings of poor decision-making on neuropsychological tests.

The presence of cognitive abnormalities has been reported to impact methamphetamine abusers’ response to treatment.³³ These findings suggest the need to develop approaches that might improve cognition in patients who are undergoing treatment for methamphetamine abuse. The monoaminergic agent modafinil and similar drugs need to be evaluated in large populations to increase the possibility of identifying characteristics of patients who might respond to cognitive enhancement.³⁴

Methamphetamine-induced psychosis. First-generation antipsychotics, such as haloperidol or fluphenazine, need to be used sparingly in patients with methamphetamine-induced psychosis because of the risk of developing extrapyramidal symptoms (EPS) and because these patients are prone to develop motor complications as a result of methamphetamine abuse. Second-generation antipsychotics, such as risperidone and olanzapine, may be more appropriate because of the lower risks of EPS.³⁵ The presence of high norepinephrine levels in some patients with recurrent methamphetamine psychosis suggests that drugs that block norepinephrine receptors, such as prazosin or propranolol, might be of therapeutic benefit if they are shown to be effective in controlled clinical trials.

The proportion of older adults in the world population is growing rapidly. In the next 10 to 15 years, the population age >60 will grow 3.5 times more rapidly than the general population.¹ As a result, there is an increased urgency in examining benefits vs risks of antipsychotics in older individuals. In a 2010 U.S. nationally representative observational study, antipsychotic use was observed to rise slowly during early and middle adulthood, peaking at approximately age 55, declining slightly between ages 55 and 65, and then rising again after age 65, with >2% of individuals ages 80 to 84 receiving an anti­psychotic.² This is likely due to the chronology of psychotic, mood, and neurocognitive disorders across the life span. In this large national study, long-term antipsychotic treatment was common, and older patients were more likely to receive their prescriptions from non-psychiatrist physicians than from psychiatrists.² Among patients receiving an antipsychotic, the proportion of those receiving it for >120 days was 54% for individuals ages 70 to 74; 49% for individuals ages 75 to 79; and 46% for individuals ages 80 to 84.

This 3-part review summarizes findings and risk–benefit considerations when prescribing antipsychotics to older individuals. Part 1 focused on those with chronic psychotic disorders, such as schizophrenia or bipolar disorder,³ and part 3 will cover patients with dementia. This review (part 2) aims to:

• briefly summarize the results of randomized controlled trials (RCTs) of second-generation antipsychotics (SGAs) and other major studies and analyses in older patients with major depressive disorder (MDD)
• provide a summative opinion on the relative risks and benefits associated with using antipsychotics in older adults with MDD
• highlight the gaps in the evidence base and areas that need additional research.

Summary of benefits, place in treatment armamentarium

The prevalence of MDD and clinically significant depressive symptoms in community­dwelling older adults is 3% to 4% and 15%, respectively, and as high as 16% and 50%, respectively, in nursing home residents.⁴ Because late-life depression is associated with suffering, disability, and excessive mortality, it needs to be recognized and treated aggressively.⁵ Antidepressants are the mainstay of pharmacotherapy for late-life depression. Guidelines and expert opinion informed by the current evidence recommend using selective serotonin reuptake inhibitors, such as escitalopram or sertraline, as a first-line treatment; serotonin norepinephrine reuptake inhibitors, such as duloxetine or venlafaxine, as a second-line treatment; and other antidepressants, such as bupropion or nortriptyline, as a third-line treatment.^5,6 However, antipsychotics also have a role in treating late-life depression.

Over the past decade, several anti­psychotics have been FDA-approved for treating MDD: aripiprazole and brexpiprazole as adjunctive treatment of MDD in adults; olanzapine-fluoxetine combination for acute and maintenance treatment of treatment-resistant depression in adults and geriatric adults; and quetiapine extended-release (XR) as monotherapy for MDD in adults and as adjunctive treatment of MDD in adults and geriatric adults who have had an inadequate response to antidepressants alone (Table 1). However, “black-box” warnings for all first-generation antipsychotics (FGAs) and SGAs alert clinicians that these medications have been associated with serious adverse events in older adults with dementia, including “deaths […] due to heart-related events (eg, heart failure, sudden death) or infections (mostly pneumonia),” with 15 of 17 placebo-controlled trials showing a higher number of deaths with an antipsychotic compared with placebo.⁷ Although similar controlled data on the mortality risk of antipsychotics in older adults with mood disorders do not exist, most experts limit their use to 2 groups of older patients: those with MDD and psychotic features (“psychotic depression”) and those with treatment-resistant depression.

Clinical trials

Olanzapine plus sertraline as first-line pharmacotherapy for MDD with psychotic features. Meyers et al¹¹ reported on a double-blind randomized comparison of olanzapine plus placebo vs olanzapine plus sertraline in 259 patients with MDD with psychotic features. An unusual feature of this trial is that it included a similar number of younger and older participants (ages 18 to 93): 117 participants were age <60 (mean age [standard deviation (SD)]: 41.3 [10.8]) and 142 were age ≥60 (mean age [SD]: 71.7 [7.8]). The same dose titration schedules based on efficacy and tolerability were used in both younger and older participants. At the end of the study, the mean dose (SD) of sertraline (or placebo) did not differ significantly in younger (174.3 mg/d [34.1]) and older participants (165.7 mg/d [43.4]). However, the mean dose (SD) of olanzapine was significantly higher in younger patients (15.7 mg/d [4.7]) than in older participants (13.4 mg/d [5.1]).

In both age groups, olanzapine plus sertraline was more efficacious than olanzapine plus placebo, and there was no statistical interaction between age, time, and treatment group (ie, the trajectories of improvement were similar in older and younger patients receiving either olanzapine or olanzapine plus sertraline). Similarly, drop-out rates because of poor tolerability did not differ significantly in younger (4.3%) and older participants (5.6%). However, in a multinomial regression, older participants were more likely to discontinue treatment because of poor tolerability.²² Older participants were significantly less likely to experience weight gain (mean [SD]: +3.3 [4.9] vs +6.5 [6.6] kg) or an increase in fasting glucose and more likely to experience a fall, pedal edema, or extrapyramidal symptoms.^11,22-24 Cholesterol and triglyceride increased significantly and similarly in both age groups. The incidence of symptoms of tardive dyskinesia (TD) over the 12-week trial was low (<5%) in both younger and older participants, and clinically diagnosed TD was reported in only 1 (older) participant.²⁵

Venlafaxine plus aripiprazole for treatment-resistant MDD. In the largest double-blind randomized study of augmentation pharmacotherapy for late-life treatment-resistant depression published to date, Lenze et al²¹ compared venlafaxine plus aripiprazole vs venlafaxine plus placebo in 181 patients age >60 (mean age 66, with 49 participants age >70) with MDD who did not remit after 12 weeks of treatment with venlafaxine (up to 300 mg/d). After 12 weeks of augmentation, remission rates were significantly higher with aripiprazole than with placebo: 40 (44%) vs 26 (29%); odds ratio (95% confidence interval [CI]): 2.0 (1.1 to 3.7). The median final aripiprazole dose was 7 mg/d (range 2 to 15 mg/d) in remitters and 10 mg/d (range 2 to 15 mg/d) in nonremitters.

Five of 90 participants (5%) discontinued aripiprazole (1 each: suicide, jitteriness/akathisia, worsening parkinsonism; and 2 withdrew consent); 8 of 90 (9%) discontinued placebo (2 each: lack of efficacy, headache; 1: worsening parkinsonism; and 3 withdrew consent). The completed suicide occurred after 5 weeks of treatment with aripiprazole and was judged to be “neither due to emergent suicidal ideation nor to aripiprazole side-effects, but was concluded by investigators to be a result of the individual’s persisting and long-standing suicidal ideation.”²¹ Including the suicide, there were 4 serious adverse events (5%) in those receiving aripiprazole (1 each: suicide, congestive heart failure, mild stroke, and diverticulitis) and 2 (2%) in those receiving placebo (1 each: myocardial infarction, hospitalized for vomiting due to accidentally taking extra venlafaxine). In 86 participants receiving aripiprazole and 87 receiving placebo, the most frequently reported adverse effects were increased dream activity (aripiprazole: 23 [27%] vs placebo: 12 [14%]), weight gain (17 [20%] vs 8 [9%]), and tremor (5 [6%] vs 0). Akathisia and parkinsonism were observed more frequently with aripiprazole than with placebo (akathisia: 24 [26%] of 91 vs 11 [12%] of 90; parkinsonism: 15 [17%] of 86 vs 2 [2%] of 81). Akathisia was generally mild and resolved with dose adjustment; however, it was associated with a transient increase in suicidality in 3 (3%) participants receiving aripiprazole vs 0 receiving placebo and persisted at the end of the trial in 5 (5%) participants receiving aripiprazole vs 2 (2%) receiving placebo. Participants receiving aripiprazole had a significantly larger increase in weight (mean [SD]: +1.93 [3.00] vs +0.01 [3.15] kg), but there were no differences between aripiprazole and placebo in changes in body fat, total cholesterol, high-density lipoprotein, low-density lipoprotein, triglycerides, glucose, insulin concentration, or QTc.

Citalopram plus risperidone for treatment-resistant MDD. Alexopoulos et al²⁶ reported an analysis of data from 110 patients age ≥55 years (mean age [SD]: 63.4 [4.8]), among 489 mixed-age patients with MDD. Participants (n = 110) who did not respond to 1 to 3 antidepressants (venlafaxine, sertraline, mirtazapine, fluoxetine, paroxetine, or bupropion in >90%) during their current depressive episode completed 4 to 6 weeks of treatment with citalopram up to 40 mg/d; 93 did not respond and were treated with open-label risperidone (0.25 to 1 mg/d) augmentation for 4 to 6 weeks. Sixty-three (68%) of these 93 patients remitted and were randomized to 24 weeks of double-blind continuation treatment with citalopram plus risperidone vs citalopram plus placebo. Neither the median times to relapse (105 vs 57 days) nor the relapse rates (risperidone: 18 of 32 [56%] vs placebo: 20 of 31 [65%]) differed significantly. During the open-label risperidone augmentation, the most common adverse events were dizziness and dry mouth (n = 9 each, 9.7% of 93). During the continuation phase, headache (n = 3; 9.1% of 32) was observed with risperidone but not with placebo (n = 0). There was no incident parkinsonism or abnormal movements noted, but risperidone was associated with weight gain during both the open-label risperidone augmentation phase (mean [SD]: +0.9 [2.1] kg) and the continuation phase (risperidone: +0.8 [3.5] vs placebo: −0.3 [2.8] kg).

Quetiapine XR monotherapy for MDD. Katila et al²⁷ reported on a placebo-controlled RCT of quetiapine XR (median dose, 158.7 mg/d; range, 50 to 300 mg/d) in 338 patients age ≥66 years (mean age [SD], 71.3 [7.5]) presenting with MDD and a major depressive episode with a duration <1 year and no history of failed antidepressants trials from 2 classes (more than two-thirds of participants had not received treatment). After 9 weeks, the reduction in depressive symptoms on the Montgomery-Åsberg Depression Rating Scale was significantly larger with quetiapine XR than with placebo (mean [SD]: −16.0 [9.3] vs −9.0 [9.9]). There were congruent, significant differences between quetiapine and placebo in terms of response rate (quetiapine XR: 105 of 164 [64%] vs placebo: 52 of 171 [30.4%]) and remission rate (92 of 164 [56.1%] vs placebo: 40 of 171 [23.4%]). The drop-out rates for all causes were similar, but the drop-out rate attributed to adverse events was higher with quetiapine than placebo (16 of 166 [9.6%] vs 7 of 172 [4.1%]). Most quetiapine drop-outs were attributable to dizziness, headache, and somnolence (n = 4 each), and placebo drop-outs were because of headache (n = 2). Consistent with the profile of quetiapine, adverse events with a rate that was at least 5% higher with quetiapine than with placebo included somnolence (64 of 166 [38.6%] vs 16 of 172 [9.3%]), dry mouth (34 [20.5%] vs 18 [10.5%]), and extrapyramidal symptoms (12 [7.2%] vs 4 [2.3%]). Changes in weight and laboratory test results (eg, glucose, lipid profile) were minimal and not clinically meaningful.

Other clinical data. The efficacy and relatively good tolerability of aripiprazole in older patients with treatment-resistant depression observed in the RCT by Lenze et al²¹ is congruent with the earlier results of 2 small (N = 20 and 24) pilot studies.^18,19 In both studies, the remission rate was 50%, and the most prevalent adverse effects were agitation/restlessness/akathisia or drowsiness/sedation. Similarly, in a post hoc pooled analysis of 409 participants ages 50 to 67 from 3 placebo-controlled randomized trials, the remission rate was significantly higher with aripiprazole than with placebo (32.5% vs 17.1%), and the most common adverse effects were akathisia or restlessness (64 of 210 [30.4%]), somnolence (18 [8.6%]), and insomnia (17 [8.1%]).²⁰

Clinical considerations

When assessing the relative benefits and risks of antipsychotics in older patients, it is important to remember that conclusions and summative opinions are necessarily influenced by the source of the data. Because much of what we know about the use of antipsychotics in geriatric adults is from clinical trials, we know more about their acute efficacy and tolerability than their long-term effectiveness and safety.²⁸ There are similar issues regarding the role of antipsychotics in treating MDD in late life. Based on the results of several RCTs,^8,11 a combination of an antidepressant plus an antipsychotic is the recommended pharmacotherapy for the acute treatment of MDD with psychotic features (Table 2).^{8,11,12,19-21,23-27} However, there are no published data to guide how long the antipsychotic should be continued.²⁹

In older patients with MDD without psychotic features, 1 relatively large placebo-controlled RCT,²¹ 2 smaller open studies,^18,19 and a post hoc analysis of a large placebo-controlled RCT in mixed-age adults²⁰ support the efficacy and relatively good tolerability of aripiprazole augmentation of an antidepressant for treatment-resistant MDD. Similarly, 1 large placebo-controlled RCT supports the efficacy and relatively good tolerability of quetiapine for non–treatment-resistant MDD. However, there are no comparative data assessing the relative merits of using these antipsy­chotics vs other pharmacologic strategies (eg, switching to another antidepressant, lithium augmentation, or combination of 2 antidepressants). Because older patients are more likely to experience adverse effects that may have more serious consequences (Table 3), many prudent clinicians reserve using antipsychotics as a third-line treatment in older patients with MDD without psychotic features and limit the duration of their use to a few months.³⁰

The proportion of older adults in the world population is growing rapidly. In the next 10 to 15 years, the population age >60 will grow 3.5 times more rapidly than the general population.¹ As a result, there is an increased urgency in examining benefits vs risks of antipsychotics in older individuals. In a 2010 U.S. nationally representative observational study, antipsychotic use was observed to rise slowly during early and middle adulthood, peaking at approximately age 55, declining slightly between ages 55 and 65, and then rising again after age 65, with >2% of individuals ages 80 to 84 receiving an anti­psychotic.² This is likely due to the chronology of psychotic, mood, and neurocognitive disorders across the life span. In this large national study, long-term antipsychotic treatment was common, and older patients were more likely to receive their prescriptions from non-psychiatrist physicians than from psychiatrists.² Among patients receiving an antipsychotic, the proportion of those receiving it for >120 days was 54% for individuals ages 70 to 74; 49% for individuals ages 75 to 79; and 46% for individuals ages 80 to 84.

This 3-part review summarizes findings and risk–benefit considerations when prescribing antipsychotics to older individuals. Part 1 focused on those with chronic psychotic disorders, such as schizophrenia or bipolar disorder,³ and part 3 will cover patients with dementia. This review (part 2) aims to:

• briefly summarize the results of randomized controlled trials (RCTs) of second-generation antipsychotics (SGAs) and other major studies and analyses in older patients with major depressive disorder (MDD)
• provide a summative opinion on the relative risks and benefits associated with using antipsychotics in older adults with MDD
• highlight the gaps in the evidence base and areas that need additional research.

Summary of benefits, place in treatment armamentarium

The prevalence of MDD and clinically significant depressive symptoms in community­dwelling older adults is 3% to 4% and 15%, respectively, and as high as 16% and 50%, respectively, in nursing home residents.⁴ Because late-life depression is associated with suffering, disability, and excessive mortality, it needs to be recognized and treated aggressively.⁵ Antidepressants are the mainstay of pharmacotherapy for late-life depression. Guidelines and expert opinion informed by the current evidence recommend using selective serotonin reuptake inhibitors, such as escitalopram or sertraline, as a first-line treatment; serotonin norepinephrine reuptake inhibitors, such as duloxetine or venlafaxine, as a second-line treatment; and other antidepressants, such as bupropion or nortriptyline, as a third-line treatment.^5,6 However, antipsychotics also have a role in treating late-life depression.

Over the past decade, several anti­psychotics have been FDA-approved for treating MDD: aripiprazole and brexpiprazole as adjunctive treatment of MDD in adults; olanzapine-fluoxetine combination for acute and maintenance treatment of treatment-resistant depression in adults and geriatric adults; and quetiapine extended-release (XR) as monotherapy for MDD in adults and as adjunctive treatment of MDD in adults and geriatric adults who have had an inadequate response to antidepressants alone (Table 1). However, “black-box” warnings for all first-generation antipsychotics (FGAs) and SGAs alert clinicians that these medications have been associated with serious adverse events in older adults with dementia, including “deaths […] due to heart-related events (eg, heart failure, sudden death) or infections (mostly pneumonia),” with 15 of 17 placebo-controlled trials showing a higher number of deaths with an antipsychotic compared with placebo.⁷ Although similar controlled data on the mortality risk of antipsychotics in older adults with mood disorders do not exist, most experts limit their use to 2 groups of older patients: those with MDD and psychotic features (“psychotic depression”) and those with treatment-resistant depression.

Clinical trials

Olanzapine plus sertraline as first-line pharmacotherapy for MDD with psychotic features. Meyers et al¹¹ reported on a double-blind randomized comparison of olanzapine plus placebo vs olanzapine plus sertraline in 259 patients with MDD with psychotic features. An unusual feature of this trial is that it included a similar number of younger and older participants (ages 18 to 93): 117 participants were age <60 (mean age [standard deviation (SD)]: 41.3 [10.8]) and 142 were age ≥60 (mean age [SD]: 71.7 [7.8]). The same dose titration schedules based on efficacy and tolerability were used in both younger and older participants. At the end of the study, the mean dose (SD) of sertraline (or placebo) did not differ significantly in younger (174.3 mg/d [34.1]) and older participants (165.7 mg/d [43.4]). However, the mean dose (SD) of olanzapine was significantly higher in younger patients (15.7 mg/d [4.7]) than in older participants (13.4 mg/d [5.1]).

In both age groups, olanzapine plus sertraline was more efficacious than olanzapine plus placebo, and there was no statistical interaction between age, time, and treatment group (ie, the trajectories of improvement were similar in older and younger patients receiving either olanzapine or olanzapine plus sertraline). Similarly, drop-out rates because of poor tolerability did not differ significantly in younger (4.3%) and older participants (5.6%). However, in a multinomial regression, older participants were more likely to discontinue treatment because of poor tolerability.²² Older participants were significantly less likely to experience weight gain (mean [SD]: +3.3 [4.9] vs +6.5 [6.6] kg) or an increase in fasting glucose and more likely to experience a fall, pedal edema, or extrapyramidal symptoms.^11,22-24 Cholesterol and triglyceride increased significantly and similarly in both age groups. The incidence of symptoms of tardive dyskinesia (TD) over the 12-week trial was low (<5%) in both younger and older participants, and clinically diagnosed TD was reported in only 1 (older) participant.²⁵

Venlafaxine plus aripiprazole for treatment-resistant MDD. In the largest double-blind randomized study of augmentation pharmacotherapy for late-life treatment-resistant depression published to date, Lenze et al²¹ compared venlafaxine plus aripiprazole vs venlafaxine plus placebo in 181 patients age >60 (mean age 66, with 49 participants age >70) with MDD who did not remit after 12 weeks of treatment with venlafaxine (up to 300 mg/d). After 12 weeks of augmentation, remission rates were significantly higher with aripiprazole than with placebo: 40 (44%) vs 26 (29%); odds ratio (95% confidence interval [CI]): 2.0 (1.1 to 3.7). The median final aripiprazole dose was 7 mg/d (range 2 to 15 mg/d) in remitters and 10 mg/d (range 2 to 15 mg/d) in nonremitters.

Five of 90 participants (5%) discontinued aripiprazole (1 each: suicide, jitteriness/akathisia, worsening parkinsonism; and 2 withdrew consent); 8 of 90 (9%) discontinued placebo (2 each: lack of efficacy, headache; 1: worsening parkinsonism; and 3 withdrew consent). The completed suicide occurred after 5 weeks of treatment with aripiprazole and was judged to be “neither due to emergent suicidal ideation nor to aripiprazole side-effects, but was concluded by investigators to be a result of the individual’s persisting and long-standing suicidal ideation.”²¹ Including the suicide, there were 4 serious adverse events (5%) in those receiving aripiprazole (1 each: suicide, congestive heart failure, mild stroke, and diverticulitis) and 2 (2%) in those receiving placebo (1 each: myocardial infarction, hospitalized for vomiting due to accidentally taking extra venlafaxine). In 86 participants receiving aripiprazole and 87 receiving placebo, the most frequently reported adverse effects were increased dream activity (aripiprazole: 23 [27%] vs placebo: 12 [14%]), weight gain (17 [20%] vs 8 [9%]), and tremor (5 [6%] vs 0). Akathisia and parkinsonism were observed more frequently with aripiprazole than with placebo (akathisia: 24 [26%] of 91 vs 11 [12%] of 90; parkinsonism: 15 [17%] of 86 vs 2 [2%] of 81). Akathisia was generally mild and resolved with dose adjustment; however, it was associated with a transient increase in suicidality in 3 (3%) participants receiving aripiprazole vs 0 receiving placebo and persisted at the end of the trial in 5 (5%) participants receiving aripiprazole vs 2 (2%) receiving placebo. Participants receiving aripiprazole had a significantly larger increase in weight (mean [SD]: +1.93 [3.00] vs +0.01 [3.15] kg), but there were no differences between aripiprazole and placebo in changes in body fat, total cholesterol, high-density lipoprotein, low-density lipoprotein, triglycerides, glucose, insulin concentration, or QTc.

Citalopram plus risperidone for treatment-resistant MDD. Alexopoulos et al²⁶ reported an analysis of data from 110 patients age ≥55 years (mean age [SD]: 63.4 [4.8]), among 489 mixed-age patients with MDD. Participants (n = 110) who did not respond to 1 to 3 antidepressants (venlafaxine, sertraline, mirtazapine, fluoxetine, paroxetine, or bupropion in >90%) during their current depressive episode completed 4 to 6 weeks of treatment with citalopram up to 40 mg/d; 93 did not respond and were treated with open-label risperidone (0.25 to 1 mg/d) augmentation for 4 to 6 weeks. Sixty-three (68%) of these 93 patients remitted and were randomized to 24 weeks of double-blind continuation treatment with citalopram plus risperidone vs citalopram plus placebo. Neither the median times to relapse (105 vs 57 days) nor the relapse rates (risperidone: 18 of 32 [56%] vs placebo: 20 of 31 [65%]) differed significantly. During the open-label risperidone augmentation, the most common adverse events were dizziness and dry mouth (n = 9 each, 9.7% of 93). During the continuation phase, headache (n = 3; 9.1% of 32) was observed with risperidone but not with placebo (n = 0). There was no incident parkinsonism or abnormal movements noted, but risperidone was associated with weight gain during both the open-label risperidone augmentation phase (mean [SD]: +0.9 [2.1] kg) and the continuation phase (risperidone: +0.8 [3.5] vs placebo: −0.3 [2.8] kg).

Quetiapine XR monotherapy for MDD. Katila et al²⁷ reported on a placebo-controlled RCT of quetiapine XR (median dose, 158.7 mg/d; range, 50 to 300 mg/d) in 338 patients age ≥66 years (mean age [SD], 71.3 [7.5]) presenting with MDD and a major depressive episode with a duration <1 year and no history of failed antidepressants trials from 2 classes (more than two-thirds of participants had not received treatment). After 9 weeks, the reduction in depressive symptoms on the Montgomery-Åsberg Depression Rating Scale was significantly larger with quetiapine XR than with placebo (mean [SD]: −16.0 [9.3] vs −9.0 [9.9]). There were congruent, significant differences between quetiapine and placebo in terms of response rate (quetiapine XR: 105 of 164 [64%] vs placebo: 52 of 171 [30.4%]) and remission rate (92 of 164 [56.1%] vs placebo: 40 of 171 [23.4%]). The drop-out rates for all causes were similar, but the drop-out rate attributed to adverse events was higher with quetiapine than placebo (16 of 166 [9.6%] vs 7 of 172 [4.1%]). Most quetiapine drop-outs were attributable to dizziness, headache, and somnolence (n = 4 each), and placebo drop-outs were because of headache (n = 2). Consistent with the profile of quetiapine, adverse events with a rate that was at least 5% higher with quetiapine than with placebo included somnolence (64 of 166 [38.6%] vs 16 of 172 [9.3%]), dry mouth (34 [20.5%] vs 18 [10.5%]), and extrapyramidal symptoms (12 [7.2%] vs 4 [2.3%]). Changes in weight and laboratory test results (eg, glucose, lipid profile) were minimal and not clinically meaningful.

Other clinical data. The efficacy and relatively good tolerability of aripiprazole in older patients with treatment-resistant depression observed in the RCT by Lenze et al²¹ is congruent with the earlier results of 2 small (N = 20 and 24) pilot studies.^18,19 In both studies, the remission rate was 50%, and the most prevalent adverse effects were agitation/restlessness/akathisia or drowsiness/sedation. Similarly, in a post hoc pooled analysis of 409 participants ages 50 to 67 from 3 placebo-controlled randomized trials, the remission rate was significantly higher with aripiprazole than with placebo (32.5% vs 17.1%), and the most common adverse effects were akathisia or restlessness (64 of 210 [30.4%]), somnolence (18 [8.6%]), and insomnia (17 [8.1%]).²⁰

Clinical considerations

When assessing the relative benefits and risks of antipsychotics in older patients, it is important to remember that conclusions and summative opinions are necessarily influenced by the source of the data. Because much of what we know about the use of antipsychotics in geriatric adults is from clinical trials, we know more about their acute efficacy and tolerability than their long-term effectiveness and safety.²⁸ There are similar issues regarding the role of antipsychotics in treating MDD in late life. Based on the results of several RCTs,^8,11 a combination of an antidepressant plus an antipsychotic is the recommended pharmacotherapy for the acute treatment of MDD with psychotic features (Table 2).^{8,11,12,19-21,23-27} However, there are no published data to guide how long the antipsychotic should be continued.²⁹

In older patients with MDD without psychotic features, 1 relatively large placebo-controlled RCT,²¹ 2 smaller open studies,^18,19 and a post hoc analysis of a large placebo-controlled RCT in mixed-age adults²⁰ support the efficacy and relatively good tolerability of aripiprazole augmentation of an antidepressant for treatment-resistant MDD. Similarly, 1 large placebo-controlled RCT supports the efficacy and relatively good tolerability of quetiapine for non–treatment-resistant MDD. However, there are no comparative data assessing the relative merits of using these antipsy­chotics vs other pharmacologic strategies (eg, switching to another antidepressant, lithium augmentation, or combination of 2 antidepressants). Because older patients are more likely to experience adverse effects that may have more serious consequences (Table 3), many prudent clinicians reserve using antipsychotics as a third-line treatment in older patients with MDD without psychotic features and limit the duration of their use to a few months.³⁰

The proportion of older adults in the world population is growing rapidly. In the next 10 to 15 years, the population age >60 will grow 3.5 times more rapidly than the general population.¹ As a result, there is an increased urgency in examining benefits vs risks of antipsychotics in older individuals. In a 2010 U.S. nationally representative observational study, antipsychotic use was observed to rise slowly during early and middle adulthood, peaking at approximately age 55, declining slightly between ages 55 and 65, and then rising again after age 65, with >2% of individuals ages 80 to 84 receiving an anti­psychotic.² This is likely due to the chronology of psychotic, mood, and neurocognitive disorders across the life span. In this large national study, long-term antipsychotic treatment was common, and older patients were more likely to receive their prescriptions from non-psychiatrist physicians than from psychiatrists.² Among patients receiving an antipsychotic, the proportion of those receiving it for >120 days was 54% for individuals ages 70 to 74; 49% for individuals ages 75 to 79; and 46% for individuals ages 80 to 84.

This 3-part review summarizes findings and risk–benefit considerations when prescribing antipsychotics to older individuals. Part 1 focused on those with chronic psychotic disorders, such as schizophrenia or bipolar disorder,³ and part 3 will cover patients with dementia. This review (part 2) aims to:

• briefly summarize the results of randomized controlled trials (RCTs) of second-generation antipsychotics (SGAs) and other major studies and analyses in older patients with major depressive disorder (MDD)
• provide a summative opinion on the relative risks and benefits associated with using antipsychotics in older adults with MDD
• highlight the gaps in the evidence base and areas that need additional research.

Summary of benefits, place in treatment armamentarium

The prevalence of MDD and clinically significant depressive symptoms in community­dwelling older adults is 3% to 4% and 15%, respectively, and as high as 16% and 50%, respectively, in nursing home residents.⁴ Because late-life depression is associated with suffering, disability, and excessive mortality, it needs to be recognized and treated aggressively.⁵ Antidepressants are the mainstay of pharmacotherapy for late-life depression. Guidelines and expert opinion informed by the current evidence recommend using selective serotonin reuptake inhibitors, such as escitalopram or sertraline, as a first-line treatment; serotonin norepinephrine reuptake inhibitors, such as duloxetine or venlafaxine, as a second-line treatment; and other antidepressants, such as bupropion or nortriptyline, as a third-line treatment.^5,6 However, antipsychotics also have a role in treating late-life depression.

Over the past decade, several anti­psychotics have been FDA-approved for treating MDD: aripiprazole and brexpiprazole as adjunctive treatment of MDD in adults; olanzapine-fluoxetine combination for acute and maintenance treatment of treatment-resistant depression in adults and geriatric adults; and quetiapine extended-release (XR) as monotherapy for MDD in adults and as adjunctive treatment of MDD in adults and geriatric adults who have had an inadequate response to antidepressants alone (Table 1). However, “black-box” warnings for all first-generation antipsychotics (FGAs) and SGAs alert clinicians that these medications have been associated with serious adverse events in older adults with dementia, including “deaths […] due to heart-related events (eg, heart failure, sudden death) or infections (mostly pneumonia),” with 15 of 17 placebo-controlled trials showing a higher number of deaths with an antipsychotic compared with placebo.⁷ Although similar controlled data on the mortality risk of antipsychotics in older adults with mood disorders do not exist, most experts limit their use to 2 groups of older patients: those with MDD and psychotic features (“psychotic depression”) and those with treatment-resistant depression.

Clinical trials

Olanzapine plus sertraline as first-line pharmacotherapy for MDD with psychotic features. Meyers et al¹¹ reported on a double-blind randomized comparison of olanzapine plus placebo vs olanzapine plus sertraline in 259 patients with MDD with psychotic features. An unusual feature of this trial is that it included a similar number of younger and older participants (ages 18 to 93): 117 participants were age <60 (mean age [standard deviation (SD)]: 41.3 [10.8]) and 142 were age ≥60 (mean age [SD]: 71.7 [7.8]). The same dose titration schedules based on efficacy and tolerability were used in both younger and older participants. At the end of the study, the mean dose (SD) of sertraline (or placebo) did not differ significantly in younger (174.3 mg/d [34.1]) and older participants (165.7 mg/d [43.4]). However, the mean dose (SD) of olanzapine was significantly higher in younger patients (15.7 mg/d [4.7]) than in older participants (13.4 mg/d [5.1]).

In both age groups, olanzapine plus sertraline was more efficacious than olanzapine plus placebo, and there was no statistical interaction between age, time, and treatment group (ie, the trajectories of improvement were similar in older and younger patients receiving either olanzapine or olanzapine plus sertraline). Similarly, drop-out rates because of poor tolerability did not differ significantly in younger (4.3%) and older participants (5.6%). However, in a multinomial regression, older participants were more likely to discontinue treatment because of poor tolerability.²² Older participants were significantly less likely to experience weight gain (mean [SD]: +3.3 [4.9] vs +6.5 [6.6] kg) or an increase in fasting glucose and more likely to experience a fall, pedal edema, or extrapyramidal symptoms.^11,22-24 Cholesterol and triglyceride increased significantly and similarly in both age groups. The incidence of symptoms of tardive dyskinesia (TD) over the 12-week trial was low (<5%) in both younger and older participants, and clinically diagnosed TD was reported in only 1 (older) participant.²⁵

Venlafaxine plus aripiprazole for treatment-resistant MDD. In the largest double-blind randomized study of augmentation pharmacotherapy for late-life treatment-resistant depression published to date, Lenze et al²¹ compared venlafaxine plus aripiprazole vs venlafaxine plus placebo in 181 patients age >60 (mean age 66, with 49 participants age >70) with MDD who did not remit after 12 weeks of treatment with venlafaxine (up to 300 mg/d). After 12 weeks of augmentation, remission rates were significantly higher with aripiprazole than with placebo: 40 (44%) vs 26 (29%); odds ratio (95% confidence interval [CI]): 2.0 (1.1 to 3.7). The median final aripiprazole dose was 7 mg/d (range 2 to 15 mg/d) in remitters and 10 mg/d (range 2 to 15 mg/d) in nonremitters.

Five of 90 participants (5%) discontinued aripiprazole (1 each: suicide, jitteriness/akathisia, worsening parkinsonism; and 2 withdrew consent); 8 of 90 (9%) discontinued placebo (2 each: lack of efficacy, headache; 1: worsening parkinsonism; and 3 withdrew consent). The completed suicide occurred after 5 weeks of treatment with aripiprazole and was judged to be “neither due to emergent suicidal ideation nor to aripiprazole side-effects, but was concluded by investigators to be a result of the individual’s persisting and long-standing suicidal ideation.”²¹ Including the suicide, there were 4 serious adverse events (5%) in those receiving aripiprazole (1 each: suicide, congestive heart failure, mild stroke, and diverticulitis) and 2 (2%) in those receiving placebo (1 each: myocardial infarction, hospitalized for vomiting due to accidentally taking extra venlafaxine). In 86 participants receiving aripiprazole and 87 receiving placebo, the most frequently reported adverse effects were increased dream activity (aripiprazole: 23 [27%] vs placebo: 12 [14%]), weight gain (17 [20%] vs 8 [9%]), and tremor (5 [6%] vs 0). Akathisia and parkinsonism were observed more frequently with aripiprazole than with placebo (akathisia: 24 [26%] of 91 vs 11 [12%] of 90; parkinsonism: 15 [17%] of 86 vs 2 [2%] of 81). Akathisia was generally mild and resolved with dose adjustment; however, it was associated with a transient increase in suicidality in 3 (3%) participants receiving aripiprazole vs 0 receiving placebo and persisted at the end of the trial in 5 (5%) participants receiving aripiprazole vs 2 (2%) receiving placebo. Participants receiving aripiprazole had a significantly larger increase in weight (mean [SD]: +1.93 [3.00] vs +0.01 [3.15] kg), but there were no differences between aripiprazole and placebo in changes in body fat, total cholesterol, high-density lipoprotein, low-density lipoprotein, triglycerides, glucose, insulin concentration, or QTc.

Citalopram plus risperidone for treatment-resistant MDD. Alexopoulos et al²⁶ reported an analysis of data from 110 patients age ≥55 years (mean age [SD]: 63.4 [4.8]), among 489 mixed-age patients with MDD. Participants (n = 110) who did not respond to 1 to 3 antidepressants (venlafaxine, sertraline, mirtazapine, fluoxetine, paroxetine, or bupropion in >90%) during their current depressive episode completed 4 to 6 weeks of treatment with citalopram up to 40 mg/d; 93 did not respond and were treated with open-label risperidone (0.25 to 1 mg/d) augmentation for 4 to 6 weeks. Sixty-three (68%) of these 93 patients remitted and were randomized to 24 weeks of double-blind continuation treatment with citalopram plus risperidone vs citalopram plus placebo. Neither the median times to relapse (105 vs 57 days) nor the relapse rates (risperidone: 18 of 32 [56%] vs placebo: 20 of 31 [65%]) differed significantly. During the open-label risperidone augmentation, the most common adverse events were dizziness and dry mouth (n = 9 each, 9.7% of 93). During the continuation phase, headache (n = 3; 9.1% of 32) was observed with risperidone but not with placebo (n = 0). There was no incident parkinsonism or abnormal movements noted, but risperidone was associated with weight gain during both the open-label risperidone augmentation phase (mean [SD]: +0.9 [2.1] kg) and the continuation phase (risperidone: +0.8 [3.5] vs placebo: −0.3 [2.8] kg).

Quetiapine XR monotherapy for MDD. Katila et al²⁷ reported on a placebo-controlled RCT of quetiapine XR (median dose, 158.7 mg/d; range, 50 to 300 mg/d) in 338 patients age ≥66 years (mean age [SD], 71.3 [7.5]) presenting with MDD and a major depressive episode with a duration <1 year and no history of failed antidepressants trials from 2 classes (more than two-thirds of participants had not received treatment). After 9 weeks, the reduction in depressive symptoms on the Montgomery-Åsberg Depression Rating Scale was significantly larger with quetiapine XR than with placebo (mean [SD]: −16.0 [9.3] vs −9.0 [9.9]). There were congruent, significant differences between quetiapine and placebo in terms of response rate (quetiapine XR: 105 of 164 [64%] vs placebo: 52 of 171 [30.4%]) and remission rate (92 of 164 [56.1%] vs placebo: 40 of 171 [23.4%]). The drop-out rates for all causes were similar, but the drop-out rate attributed to adverse events was higher with quetiapine than placebo (16 of 166 [9.6%] vs 7 of 172 [4.1%]). Most quetiapine drop-outs were attributable to dizziness, headache, and somnolence (n = 4 each), and placebo drop-outs were because of headache (n = 2). Consistent with the profile of quetiapine, adverse events with a rate that was at least 5% higher with quetiapine than with placebo included somnolence (64 of 166 [38.6%] vs 16 of 172 [9.3%]), dry mouth (34 [20.5%] vs 18 [10.5%]), and extrapyramidal symptoms (12 [7.2%] vs 4 [2.3%]). Changes in weight and laboratory test results (eg, glucose, lipid profile) were minimal and not clinically meaningful.

Other clinical data. The efficacy and relatively good tolerability of aripiprazole in older patients with treatment-resistant depression observed in the RCT by Lenze et al²¹ is congruent with the earlier results of 2 small (N = 20 and 24) pilot studies.^18,19 In both studies, the remission rate was 50%, and the most prevalent adverse effects were agitation/restlessness/akathisia or drowsiness/sedation. Similarly, in a post hoc pooled analysis of 409 participants ages 50 to 67 from 3 placebo-controlled randomized trials, the remission rate was significantly higher with aripiprazole than with placebo (32.5% vs 17.1%), and the most common adverse effects were akathisia or restlessness (64 of 210 [30.4%]), somnolence (18 [8.6%]), and insomnia (17 [8.1%]).²⁰

Clinical considerations

When assessing the relative benefits and risks of antipsychotics in older patients, it is important to remember that conclusions and summative opinions are necessarily influenced by the source of the data. Because much of what we know about the use of antipsychotics in geriatric adults is from clinical trials, we know more about their acute efficacy and tolerability than their long-term effectiveness and safety.²⁸ There are similar issues regarding the role of antipsychotics in treating MDD in late life. Based on the results of several RCTs,^8,11 a combination of an antidepressant plus an antipsychotic is the recommended pharmacotherapy for the acute treatment of MDD with psychotic features (Table 2).^{8,11,12,19-21,23-27} However, there are no published data to guide how long the antipsychotic should be continued.²⁹

In older patients with MDD without psychotic features, 1 relatively large placebo-controlled RCT,²¹ 2 smaller open studies,^18,19 and a post hoc analysis of a large placebo-controlled RCT in mixed-age adults²⁰ support the efficacy and relatively good tolerability of aripiprazole augmentation of an antidepressant for treatment-resistant MDD. Similarly, 1 large placebo-controlled RCT supports the efficacy and relatively good tolerability of quetiapine for non–treatment-resistant MDD. However, there are no comparative data assessing the relative merits of using these antipsy­chotics vs other pharmacologic strategies (eg, switching to another antidepressant, lithium augmentation, or combination of 2 antidepressants). Because older patients are more likely to experience adverse effects that may have more serious consequences (Table 3), many prudent clinicians reserve using antipsychotics as a third-line treatment in older patients with MDD without psychotic features and limit the duration of their use to a few months.³⁰

There are many rewards for full-time academic psychiatrists such as myself, including didactic teaching, clinical supervision, and 1:1 mentorship of freshly minted medical school graduates, transforming them into accomplished clinical psychiatrists. The technical and personal growth of psychiatric residents over 4 years of post-MD training can be amazing and very gratifying to witness.

But the road to clinical competence often is littered with mistakes. It is the duty of the clinical supervisor to convert every error into a learning opportunity to hone the skills of a future psychiatrist. Over time, fewer mistakes occur, not only because of maturity and seasoning, but also because psychiatric residents learn how to thoughtfully deliberate about their clinical decision-making to select the best treatment option for their patients.

Yet, even with exemplary training, the rigors and constraints of clinical practice inevitably lead to some unforced errors, mostly minor but sometimes consequential. Even experienced practitioners are not immune from making a mistake in the hustle and bustle of daily work (exacerbated by the time-consuming pressures of electronic health record documentation). No one is infallible, but everyone must avoid making the same mistake twice, even if mounting demands lead to “shortcuts” that may not necessarily put the patient at risk but could lead to suboptimal outcomes. But once in a while, a serious complication may ensue.

Here are some common errors of omission or commission that even competent practitioners may make in a busy clinical practice.

Rushing to a diagnosis. To arrive at a primary psychiatric diagnosis, all potential secondary causes must be ruled out. This includes systematic screening for possible drug-induced psychopathology related not only to drugs of abuse, but also to prescription medications, some of which can have serious iatrogenic effects, including depression, anxiety, mania, psychosis, or cognitive dulling. The other important cause to rule out is the possibility of a general medical condition triggering psychiatric symptoms, which requires targeted questioning about medical history, a review of organ systems, and ordering key laboratory tests.

Skipping a baseline cognitive assessment. Cognitive impairment, especially memory and executive function, is now well recognized as an important component of major psychiatric disorders, including schizophrenia, bipolar disorder, major depressive disorder, anxiety, and attention-deficit/hyperactivity disorder. A standardized cognitive battery can provide a valuable profile of brain functions. Knowing the patient’s cognitive strengths and weaknesses before initiating pharmacotherapy is essential to assess the positive or negative impact of the medications. It also can help with patients’ vocational rehabilitation, matching them with jobs compatible with their cognitive strengths.

Inaccurate differential diagnosis. Is it borderline personality or bipolar disorder? Is it schizophrenia or psychotic bipolar disorder? Is it unipolar or bipolar depression? Is it a conversion reaction or a genuine medical condition? The answers to such questions are critical, because inaccurate diagnosis can lead to a lack of improvement and prolonged suffering for patients or adverse effects that could be avoided.

Using a high dose of a medication immediately for a first-episode psychiatric disorder. One of the least patient-friendly medical decisions is to start a first-episode patient on a high dose of a medication on day 1. Gradual titration can circumvent intolerable adverse effects and help establish the lowest effective dose. Patient acceptance and adherence are far more likely if the patient’s brain is not “abruptly medicated.”

Using combination therapy right away. There are a few psychiatric conditions for which combination therapy is FDA-approved and regarded as “rational polypharmacy.” However, it always makes sense to start with 1 (primary) medication and assess its efficacy, tolerability, and safety before adding an adjunctive agent. Some patients may improve substantially with monotherapy, which is always preferable. Using drug combinations as the initial intervention can be problematic, especially if they are not evidence-based and off-label.

Selecting an obesogenic drug as first-line. Many psychotropics, such as antipsychotics, antidepressants, or mood stabilizers, often come as a class of several agents. Clinicians can select any member of the same class (such as selective serotonin reuptake inhibitors [SSRIs] or atypical antipsychotics) because they are all FDA-approved for efficacy. However, the major difference among what often are called “me too” drugs is the adverse effects profile. For many psychotropic medications, significant weight gain is one of the worst medical adverse effects, because it often leads to metabolic dysregulation (hyperglycemia, dyslipidemia, and hypertension) and increases the risk of cardiovascular disease. Many psychiatric patients become obese and have great difficulty losing weight, especially if they are sedentary and have poor eating habits.

Using benzodiazepines as a first-line treatment for anxiety. Although certainly efficacious, and rapidly so, benzodiazepines must be avoided as a first-line treatment for anxiety. The addiction potential is significant, and patients with anxiety will subsequently not respond adequately to standard anxiolytic pharmacotherapy, such as an SSRI, because the anxiolytic effect of these other medications is gradual and not as rapid or potent. Some primary care providers (PCPs) resort to using strong benzodiazepines (such as alprazolam) as first-line, and then refer the patient to a psychiatrist, who finds it quite challenging to steer the patient to an evidence-based option that is less harmful for long-term management. The residents and I have encountered such situations often, sometimes leading to complex interactions with patients who demand renewal of a high dose of a benzodiazepine that had been prescribed to them by a different clinician.

Low utilization of some efficacious agents. It is surprising how some useful pharmacotherapeutic strategies are not employed as often as they should be. This includes lithium for a manic episode; a long-acting injectable antipsychotic in the early phase of schizophrenia; clozapine for patients who failed to respond to a couple of antipsychotics or have chronic suicidal tendencies; lurasidone or quetiapine for bipolar depression (the only FDA-approved medications for this condition); or monoamine oxidase inhibitors for treatment-resistant depression. These drugs can be useful, although some require ongoing blood-level measurements and monitoring for efficacy and adverse effects.

Not recognizing tardive dyskinesia (TD) earlier. TD is one of the most serious neurologic complications of dopamine-receptor working agents (antipsychotics). FDA-approved treatments finally arrived in 2017, but the recognition of the abnormal oro-bucco-lingual or facial choreiform movements remain low (and the use of the Abnormal Involuntary Movement Scale to screen for TD has faded since second-generation antipsychotics were introduced). It is essential to identify this adverse effect early and treat it promptly to avoid its worsening and potential irreversibility.

Other errors of omission or commission include:

• Not collaborating actively with the patient’s PCP to integrate the medical care to improve the patient’s overall health, not just mental health. Collaborative care improves clinical outcomes for most patients.
• Not using available pharmacogenetics testing to provide the patient with “personalized medicine,” such as establishing if they are poor or rapid metabolizers of certain cytochrome hepatic enzymes or checking whether they are less likely to respond to antidepressant medications.
• “Lowering expectations” for patients with severe psychiatric disorders, giving them the message (verbally or nonverbally) that their condition is “hopeless” and that recovery is beyond their reach. Giving hope and trying hard to find better treatment options are the foundation of good medical practice, especially for the sickest patients.

Psychiatrists always aim to do the right thing for their patients, even when the pressures of clinical practice are intense and palpable. But sometimes, an inadvertent slip may occur in the form of an error of omission or commission. These unforced errors are rarely dangerous, but they have the potential to delay response, increase the disease burden, or complicate the illness course. Compassion may be in generous supply, but it’s not enough: We must strive to make our patient-centered, evidence-based clinical practice an error-free zone.

There are many rewards for full-time academic psychiatrists such as myself, including didactic teaching, clinical supervision, and 1:1 mentorship of freshly minted medical school graduates, transforming them into accomplished clinical psychiatrists. The technical and personal growth of psychiatric residents over 4 years of post-MD training can be amazing and very gratifying to witness.

But the road to clinical competence often is littered with mistakes. It is the duty of the clinical supervisor to convert every error into a learning opportunity to hone the skills of a future psychiatrist. Over time, fewer mistakes occur, not only because of maturity and seasoning, but also because psychiatric residents learn how to thoughtfully deliberate about their clinical decision-making to select the best treatment option for their patients.

Yet, even with exemplary training, the rigors and constraints of clinical practice inevitably lead to some unforced errors, mostly minor but sometimes consequential. Even experienced practitioners are not immune from making a mistake in the hustle and bustle of daily work (exacerbated by the time-consuming pressures of electronic health record documentation). No one is infallible, but everyone must avoid making the same mistake twice, even if mounting demands lead to “shortcuts” that may not necessarily put the patient at risk but could lead to suboptimal outcomes. But once in a while, a serious complication may ensue.

Here are some common errors of omission or commission that even competent practitioners may make in a busy clinical practice.

Rushing to a diagnosis. To arrive at a primary psychiatric diagnosis, all potential secondary causes must be ruled out. This includes systematic screening for possible drug-induced psychopathology related not only to drugs of abuse, but also to prescription medications, some of which can have serious iatrogenic effects, including depression, anxiety, mania, psychosis, or cognitive dulling. The other important cause to rule out is the possibility of a general medical condition triggering psychiatric symptoms, which requires targeted questioning about medical history, a review of organ systems, and ordering key laboratory tests.

Skipping a baseline cognitive assessment. Cognitive impairment, especially memory and executive function, is now well recognized as an important component of major psychiatric disorders, including schizophrenia, bipolar disorder, major depressive disorder, anxiety, and attention-deficit/hyperactivity disorder. A standardized cognitive battery can provide a valuable profile of brain functions. Knowing the patient’s cognitive strengths and weaknesses before initiating pharmacotherapy is essential to assess the positive or negative impact of the medications. It also can help with patients’ vocational rehabilitation, matching them with jobs compatible with their cognitive strengths.

Inaccurate differential diagnosis. Is it borderline personality or bipolar disorder? Is it schizophrenia or psychotic bipolar disorder? Is it unipolar or bipolar depression? Is it a conversion reaction or a genuine medical condition? The answers to such questions are critical, because inaccurate diagnosis can lead to a lack of improvement and prolonged suffering for patients or adverse effects that could be avoided.

Using a high dose of a medication immediately for a first-episode psychiatric disorder. One of the least patient-friendly medical decisions is to start a first-episode patient on a high dose of a medication on day 1. Gradual titration can circumvent intolerable adverse effects and help establish the lowest effective dose. Patient acceptance and adherence are far more likely if the patient’s brain is not “abruptly medicated.”

Using combination therapy right away. There are a few psychiatric conditions for which combination therapy is FDA-approved and regarded as “rational polypharmacy.” However, it always makes sense to start with 1 (primary) medication and assess its efficacy, tolerability, and safety before adding an adjunctive agent. Some patients may improve substantially with monotherapy, which is always preferable. Using drug combinations as the initial intervention can be problematic, especially if they are not evidence-based and off-label.

Selecting an obesogenic drug as first-line. Many psychotropics, such as antipsychotics, antidepressants, or mood stabilizers, often come as a class of several agents. Clinicians can select any member of the same class (such as selective serotonin reuptake inhibitors [SSRIs] or atypical antipsychotics) because they are all FDA-approved for efficacy. However, the major difference among what often are called “me too” drugs is the adverse effects profile. For many psychotropic medications, significant weight gain is one of the worst medical adverse effects, because it often leads to metabolic dysregulation (hyperglycemia, dyslipidemia, and hypertension) and increases the risk of cardiovascular disease. Many psychiatric patients become obese and have great difficulty losing weight, especially if they are sedentary and have poor eating habits.

Using benzodiazepines as a first-line treatment for anxiety. Although certainly efficacious, and rapidly so, benzodiazepines must be avoided as a first-line treatment for anxiety. The addiction potential is significant, and patients with anxiety will subsequently not respond adequately to standard anxiolytic pharmacotherapy, such as an SSRI, because the anxiolytic effect of these other medications is gradual and not as rapid or potent. Some primary care providers (PCPs) resort to using strong benzodiazepines (such as alprazolam) as first-line, and then refer the patient to a psychiatrist, who finds it quite challenging to steer the patient to an evidence-based option that is less harmful for long-term management. The residents and I have encountered such situations often, sometimes leading to complex interactions with patients who demand renewal of a high dose of a benzodiazepine that had been prescribed to them by a different clinician.

Low utilization of some efficacious agents. It is surprising how some useful pharmacotherapeutic strategies are not employed as often as they should be. This includes lithium for a manic episode; a long-acting injectable antipsychotic in the early phase of schizophrenia; clozapine for patients who failed to respond to a couple of antipsychotics or have chronic suicidal tendencies; lurasidone or quetiapine for bipolar depression (the only FDA-approved medications for this condition); or monoamine oxidase inhibitors for treatment-resistant depression. These drugs can be useful, although some require ongoing blood-level measurements and monitoring for efficacy and adverse effects.

Not recognizing tardive dyskinesia (TD) earlier. TD is one of the most serious neurologic complications of dopamine-receptor working agents (antipsychotics). FDA-approved treatments finally arrived in 2017, but the recognition of the abnormal oro-bucco-lingual or facial choreiform movements remain low (and the use of the Abnormal Involuntary Movement Scale to screen for TD has faded since second-generation antipsychotics were introduced). It is essential to identify this adverse effect early and treat it promptly to avoid its worsening and potential irreversibility.

Other errors of omission or commission include:

• Not collaborating actively with the patient’s PCP to integrate the medical care to improve the patient’s overall health, not just mental health. Collaborative care improves clinical outcomes for most patients.
• Not using available pharmacogenetics testing to provide the patient with “personalized medicine,” such as establishing if they are poor or rapid metabolizers of certain cytochrome hepatic enzymes or checking whether they are less likely to respond to antidepressant medications.
• “Lowering expectations” for patients with severe psychiatric disorders, giving them the message (verbally or nonverbally) that their condition is “hopeless” and that recovery is beyond their reach. Giving hope and trying hard to find better treatment options are the foundation of good medical practice, especially for the sickest patients.

Psychiatrists always aim to do the right thing for their patients, even when the pressures of clinical practice are intense and palpable. But sometimes, an inadvertent slip may occur in the form of an error of omission or commission. These unforced errors are rarely dangerous, but they have the potential to delay response, increase the disease burden, or complicate the illness course. Compassion may be in generous supply, but it’s not enough: We must strive to make our patient-centered, evidence-based clinical practice an error-free zone.

There are many rewards for full-time academic psychiatrists such as myself, including didactic teaching, clinical supervision, and 1:1 mentorship of freshly minted medical school graduates, transforming them into accomplished clinical psychiatrists. The technical and personal growth of psychiatric residents over 4 years of post-MD training can be amazing and very gratifying to witness.

But the road to clinical competence often is littered with mistakes. It is the duty of the clinical supervisor to convert every error into a learning opportunity to hone the skills of a future psychiatrist. Over time, fewer mistakes occur, not only because of maturity and seasoning, but also because psychiatric residents learn how to thoughtfully deliberate about their clinical decision-making to select the best treatment option for their patients.

Yet, even with exemplary training, the rigors and constraints of clinical practice inevitably lead to some unforced errors, mostly minor but sometimes consequential. Even experienced practitioners are not immune from making a mistake in the hustle and bustle of daily work (exacerbated by the time-consuming pressures of electronic health record documentation). No one is infallible, but everyone must avoid making the same mistake twice, even if mounting demands lead to “shortcuts” that may not necessarily put the patient at risk but could lead to suboptimal outcomes. But once in a while, a serious complication may ensue.

Here are some common errors of omission or commission that even competent practitioners may make in a busy clinical practice.

Rushing to a diagnosis. To arrive at a primary psychiatric diagnosis, all potential secondary causes must be ruled out. This includes systematic screening for possible drug-induced psychopathology related not only to drugs of abuse, but also to prescription medications, some of which can have serious iatrogenic effects, including depression, anxiety, mania, psychosis, or cognitive dulling. The other important cause to rule out is the possibility of a general medical condition triggering psychiatric symptoms, which requires targeted questioning about medical history, a review of organ systems, and ordering key laboratory tests.

Skipping a baseline cognitive assessment. Cognitive impairment, especially memory and executive function, is now well recognized as an important component of major psychiatric disorders, including schizophrenia, bipolar disorder, major depressive disorder, anxiety, and attention-deficit/hyperactivity disorder. A standardized cognitive battery can provide a valuable profile of brain functions. Knowing the patient’s cognitive strengths and weaknesses before initiating pharmacotherapy is essential to assess the positive or negative impact of the medications. It also can help with patients’ vocational rehabilitation, matching them with jobs compatible with their cognitive strengths.

Inaccurate differential diagnosis. Is it borderline personality or bipolar disorder? Is it schizophrenia or psychotic bipolar disorder? Is it unipolar or bipolar depression? Is it a conversion reaction or a genuine medical condition? The answers to such questions are critical, because inaccurate diagnosis can lead to a lack of improvement and prolonged suffering for patients or adverse effects that could be avoided.

Using a high dose of a medication immediately for a first-episode psychiatric disorder. One of the least patient-friendly medical decisions is to start a first-episode patient on a high dose of a medication on day 1. Gradual titration can circumvent intolerable adverse effects and help establish the lowest effective dose. Patient acceptance and adherence are far more likely if the patient’s brain is not “abruptly medicated.”

Using combination therapy right away. There are a few psychiatric conditions for which combination therapy is FDA-approved and regarded as “rational polypharmacy.” However, it always makes sense to start with 1 (primary) medication and assess its efficacy, tolerability, and safety before adding an adjunctive agent. Some patients may improve substantially with monotherapy, which is always preferable. Using drug combinations as the initial intervention can be problematic, especially if they are not evidence-based and off-label.

Selecting an obesogenic drug as first-line. Many psychotropics, such as antipsychotics, antidepressants, or mood stabilizers, often come as a class of several agents. Clinicians can select any member of the same class (such as selective serotonin reuptake inhibitors [SSRIs] or atypical antipsychotics) because they are all FDA-approved for efficacy. However, the major difference among what often are called “me too” drugs is the adverse effects profile. For many psychotropic medications, significant weight gain is one of the worst medical adverse effects, because it often leads to metabolic dysregulation (hyperglycemia, dyslipidemia, and hypertension) and increases the risk of cardiovascular disease. Many psychiatric patients become obese and have great difficulty losing weight, especially if they are sedentary and have poor eating habits.

Using benzodiazepines as a first-line treatment for anxiety. Although certainly efficacious, and rapidly so, benzodiazepines must be avoided as a first-line treatment for anxiety. The addiction potential is significant, and patients with anxiety will subsequently not respond adequately to standard anxiolytic pharmacotherapy, such as an SSRI, because the anxiolytic effect of these other medications is gradual and not as rapid or potent. Some primary care providers (PCPs) resort to using strong benzodiazepines (such as alprazolam) as first-line, and then refer the patient to a psychiatrist, who finds it quite challenging to steer the patient to an evidence-based option that is less harmful for long-term management. The residents and I have encountered such situations often, sometimes leading to complex interactions with patients who demand renewal of a high dose of a benzodiazepine that had been prescribed to them by a different clinician.

Low utilization of some efficacious agents. It is surprising how some useful pharmacotherapeutic strategies are not employed as often as they should be. This includes lithium for a manic episode; a long-acting injectable antipsychotic in the early phase of schizophrenia; clozapine for patients who failed to respond to a couple of antipsychotics or have chronic suicidal tendencies; lurasidone or quetiapine for bipolar depression (the only FDA-approved medications for this condition); or monoamine oxidase inhibitors for treatment-resistant depression. These drugs can be useful, although some require ongoing blood-level measurements and monitoring for efficacy and adverse effects.

Not recognizing tardive dyskinesia (TD) earlier. TD is one of the most serious neurologic complications of dopamine-receptor working agents (antipsychotics). FDA-approved treatments finally arrived in 2017, but the recognition of the abnormal oro-bucco-lingual or facial choreiform movements remain low (and the use of the Abnormal Involuntary Movement Scale to screen for TD has faded since second-generation antipsychotics were introduced). It is essential to identify this adverse effect early and treat it promptly to avoid its worsening and potential irreversibility.

Other errors of omission or commission include:

• Not collaborating actively with the patient’s PCP to integrate the medical care to improve the patient’s overall health, not just mental health. Collaborative care improves clinical outcomes for most patients.
• Not using available pharmacogenetics testing to provide the patient with “personalized medicine,” such as establishing if they are poor or rapid metabolizers of certain cytochrome hepatic enzymes or checking whether they are less likely to respond to antidepressant medications.
• “Lowering expectations” for patients with severe psychiatric disorders, giving them the message (verbally or nonverbally) that their condition is “hopeless” and that recovery is beyond their reach. Giving hope and trying hard to find better treatment options are the foundation of good medical practice, especially for the sickest patients.

Psychiatrists always aim to do the right thing for their patients, even when the pressures of clinical practice are intense and palpable. But sometimes, an inadvertent slip may occur in the form of an error of omission or commission. These unforced errors are rarely dangerous, but they have the potential to delay response, increase the disease burden, or complicate the illness course. Compassion may be in generous supply, but it’s not enough: We must strive to make our patient-centered, evidence-based clinical practice an error-free zone.

“It is a mystery how a ubiquitous treatment used since antiquity was unknown, unnamed, and unidentified until recently. It is even more remarkable because this is the only treatment common to all societies and cultures.”¹

The treatment discussed above is not a specific pill, surgery, plant, or herb. Rather, the authors are referring to placebo. Indeed, the history of medical treatment is largely a chronicle of placebos. When subjected to scientific scrutiny, the overwhelming majority of treatments have turned out to be devoid of intrinsic therapeutic value; they derived their benefits from the placebo effect. Despite these benefits, the term “placebo” comes with unfortunate baggage. Latin for “I shall please,” it is the first word of the Christian vespers for the dead. In the 12th century these vespers were commonly referred to as placebos. By the 1300s, the term had become secular and pejorative, suggesting a flatterer or sycophant. When the word entered medical terminology in the late 18th century, the negative connotation stuck. A placebo was defined as a medicine given to please patients rather than to benefit them. In the modern era, the lack of pharmacologic activity became part of the definition as well.

The word placebo brings with it connotations of deception, fakery, and ineffectiveness. But one of the things about placebos that contribute mightily to the health care community’s aversion toward them is, in fact, their effectiveness. They bring relief across a wide range of medical conditions.² In doing so, placebos impugn the value of our most cherished remedies, hamper the development of new therapeutics, and threaten our livelihoods as health professionals.³

Placebos often are conceptualized as any treatment that lacks intrinsic therapeutic value, such as sugar pills. But looking at what placebo treatment actually entails, both in placebo-controlled treatment trials and in clinical settings, suggests a more comprehensive definition. Placebos encompass all the elements common to any treatment or healing situation. These include a recognized healer, evaluation, diagnosis, prognosis, plausible treatment, and most importantly, the expectation that one will recover. Along these lines, the placebo response can be thought of as the response to the common elements of the treatment or healing situation.³

Research regarding the placebo effect has mushroomed in the past 2 decades. Over this time, we have learned a good deal about both the mechanisms underlying the placebo effect and how the placebo effect can be applied to enhance the benefit of conventional treatment. Brain imaging technology has revealed that when placebo treatment alleviates pain, Parkinson’s disease, and depression, brain changes occur that are similar to those observed with active pharma­cologic treatment.^4,5 Recent studies also show that deliberate, open (nondeceptive) use of placebo can improve the symptoms of several conditions, including depression, pain, and irritable bowel syndrome.⁶ Furthermore, intermittent substitution of placebo pills for pharmacologically active treatment in a conditioning paradigm can be as effective as the “real” treatment.⁷ Also, research over the past decade has verified that certain common features of the treatment situation, particularly the quality of the doctor–patient encounter, contribute to the placebo response and have a demonstrable impact on the outcome of treatment.⁸ Clearly, the placebo effect has gone from being simply a nuisance that interferes with the evaluation of new treatments to a variable worthy of study and application in its own right. Although, for the most part, clinical practice has not kept up with these advances.

Placebos seem to have their greatest impact on the subjective symptoms of disease—pain, distress, and discouragement. It should come as no surprise, then, that placebos are particularly effective in certain psychiatric conditions. In some forms of anxiety and depressive disorders, for example, distress is the illness, and placebos reliably bring relief. Patients with panic disorder, mild to moderate depression, or generalized anxiety disorder get almost as much relief with placebo as they do with conventional treatment (about one-half improve with placebo).^9-11 But <20% of those with obsessive-compulsive disorder improve with placebo, and placebo response rates are also low in patients with schizophrenia or dementia. Mania, attention-deficit/hyperactivity disorder (ADHD), and severe depression fall somewhere in the middle.³

Harnessing the placebo response

There may be a few circumstances in psychiatric practice when it makes sense to intentionally prescribe a placebo as treatment, and we discuss those below. But far more frequently, what we know about the elements that contribute to the placebo effect can be applied to enhance the benefits of any treatment. Patients might be best served if deliberate mobilization of the placebo effect was a standard adjunct to conventional clinical care.

Various components of the treatment situation, collectively referred to as placebo, are a powerful antidote for illness, and some of these healing components exert their influence without special activity on the clinician’s part:

• Simply seeking psychiatric care can bring relief by providing some sense of control over distressing symptoms. The standard trappings of the office or clinic and customary office procedures—from the presentation of one’s insurance card to taking a history—offer reassurance and evoke the expectation that improvement or recovery is around the corner.
• The comfort provided by the psychiatrist’s presence is enhanced when patients feel that they are in the hands of a recognized healer. Psychiatrists inspire confidence when they look like a psychiatrist, or more precisely, like the patient’s idea of what a psychiatrist should look like. In our culture, that means a white coat or business attire.

A thorough evaluation is one of the common treatment elements that does the most to reduce distress and inspire confidence. The quality of an evaluation bears a strong relationship to patients’ satisfaction with the medical encounter, and can influence the amount of disability they suffer.^3,12-15

Although guidelines for conducting effective psychiatric interviews have been around for almost 100 years, psychiatrists vary considerably in the extent to which they elicit complete and accurate information, build rapport, give patients the sense that they are listened to, and provide a thorough assessment. The degree to which patients feel that the clinician is responsive to their concerns depends as much on the style of the interview as on the amount of time devoted to it. Nonverbal behavior can carry the message that the clinician is paying full attention. Something as simple as not answering the phone during an interview (this seems obvious, but a surprising and troubling number of mental health professionals take phone calls during interviews and treatment sessions) conveys an important message about the importance that the clinician places on the patient’s problems.³

The idea that the treatment situation itself provides reassurance and reduces distress, and in doing so, powers a good bit of the placebo effect, is enshrined in such concepts as the importance of good bedside manner. Many feel that the doctor’s thoughtful attention, positive regard, and optimism—so valued by patients—are justified on humanitarian grounds alone; actual evidence that this caring behavior contributes to healing isn’t required. To many, the healing properties of the treatment situation are self-evident. But as the costs of health care snowball and the demands for efficiency and cost-effectiveness rise, the time that psychiatrists can devote to patients has dwindled. Third-party payors demand evidence, beyond intuition and common sense, that diagnostic procedures and treatments have some usefulness, and rightly so.

Is there any evidence that the common components of the treatment situation provide benefit?³ More specifically, does the quality of the doctor–patient relationship and the patient’s feelings about a therapeutic encounter promote healing? Several studies suggest that the doctor–patient relationship has a demonstrable impact on symptom relief.¹⁶ In 1 study, oncologists were randomly assigned to receive a Communication Skills Training (CST) program or not. CST included a 1.5-day face-to-face workshop and 6 hours of monthly videoconferencing that focused on improving communication skills with patients.¹⁷ Lessons included building rapport, engaging in appropriate eye contact, and normalizing difficult experiences. One week after initially consulting with their physician, patients who saw an oncologist in the CST group experienced less anxiety and depression than those who saw an oncologist who did not receive CST. The benefit of CST for patient anxiety mostly persisted at a 3-month follow-up.

A recent meta-analysis pooled the results of 47 studies to examine the relationship between how much trust patients have for their doctors and health outcomes. There was a small to medium association: More trust was associated with greater improvement.¹⁸ It is possible that a good doctor–patient relationship enhances expectancies. However, it is also likely that a positive therapeutic relationship is inherently soothing and reduces distress or dysfunction independent of expectation. Regardless of the precise mechanism, these studies warrant attention. We all understand that it is important on ethical grounds to treat patients with respect and kindness. Research shows that this type of behavior also promotes recovery.

Patient expectations. The idea that expectation of improvement has a major impact on treatment outcome is firmly grounded in research on the placebo effect. Studies have shown that what people expect to experience as an outcome of treatment has a substantial impact on what they actually experience. In a classic study, a doctor told some patients with symptoms of minor illness that they would feel better soon and another group with the same symptoms that he didn’t know what ailed them.¹⁹ Two weeks later, 64% of patients in the “positive expectation” group were improved, compared with only 39% of patients in the “negative” group. In another study, adults were exposed to an allergen that caused a skin reaction.²⁰ Hand lotion (ie, a therapeutically inert substance) was then spread on the skin. Patients were led to believe that the cream would either alleviate or exacerbate the itching. The experimentally-induced wheal-and-flare was measured in both groups a few minutes after the allergen and cream were applied. The wheal-and-flare were worse for participants in the group that expected exacerbation.

Not uncommonly, expectation can have more impact on clinical outcome than a drug’s pharmacologic activity. In a double-blind placebo-controlled study, patients with depression were treated with St. John’s wort, sertraline, or placebo.²¹ They improved to the same extent with all 3 treatments. But when patients were asked to guess the treatment to which they had been assigned, those who thought they had received placebo showed little improvement, irrespective of which intervention they actually received, and those who guessed they had been given St. John’s wort or sertraline showed uniformly large improvement, irrespective of which intervention they actually received (including placebo). The researchers concluded that “Patient beliefs regarding treatment may have a stronger association with clinical outcome than the actual medication received.”

Psychiatrists who wish to use all the therapeutic tools at their disposal must attend to and manage patient expectations. One part of channeling a patient’s expectation is to thoroughly assess the patient’s beliefs regarding the efficacy of various treatments. If a patient’s uncle said that a certain drug is a miracle cure for anxiety, and the patient believes it to be true, then that expectation must be taken into consideration. Many patients prefer alternative treatments to conventional therapies. As long as there is no reason to think an alternative treatment will cause harm, a compromise might be reasonable. For example, if a patient with schizophrenia wants to treat her symptoms with herbal tea, the psychiatrist could say, “In addition to the tea, I recommend that you also take clozapine. The combination is likely to improve your symptoms.”³ More than anything else, the words a psychiatrist uses when recommending treatment shape the patient’s expectations. “You should be feeling a lot less anxious soon after you start taking this” has a different effect than “Try this. It may help.”

Prescribing ‘open-label’ placebo

There may be some limited circumstances where an actual placebo (eg, a sugar pill) might be suitable as a treatment. These include when placebo and conventional treatment provide similar results and a patient is reluctant to take conventional medicine, or when there is no effective conventional treatment. The deceptive prescription of placebo (providing placebo and calling it a drug) has a long history and was considered ethical—and recommended by medical authorities—until the latter half of the 20th century. This practice was deemed unethical in the 1980s, because it was dishonest and violated patient autonomy. Because it was widely believed that placebos given openly would be ineffective, the end of placebo treatment seemed at hand. An intriguing body of evidence, however, suggests that placebos can be effective even when patients know they are taking a placebo. Patients given an “open-label” placebo are told something along the lines of “the pill being prescribed contains no medicine, but some people improve with it, perhaps because the pill stimulates the body’s self-healing.” Open-label placebo has been evaluated for depression,²² low back pain,²³ irritable bowel syndrome,²⁴ neurosis,²⁵ allergic rhinitis,²⁶ and anxiety.²⁷ Most of these studies are small, and some were uncontrolled. Yet they consistently have shown that symptoms improve with a nondeceptive placebo, and improve to a greater extent than with no treatment.

The most recent trial is a promising example of the potential of open-label placebos. In this study, 96 patients with chronic low back pain were randomly assigned to 3 weeks of treatment as usual (TAU) or 3 weeks of TAU plus open-label placebo.²³ Patients who received open-label placebo were educated about the placebo effect and shown a film clip describing promising results of a prior open-label placebo study. They were then given placebo pills to be take once daily, and clearly told the pills contained no active medication. After 3 weeks, patients in the TAU plus placebo group reported less pain and less disability than patients who received TAU without a placebo. Some patients even requested a placebo prescription at the end of the study.

The placebo response provides a rational basis for prescribing innocuous alternative therapies with no intrinsic therapeutic value. Patients who prefer and believe in the effectiveness of alternative remedies—herbal compounds, massage, magnets, homeopathic solutions, etc.—can be recommended these treatments to mobilize a placebo response.

Using a conditioning model. Prescribing a placebo to obtain a conditioned drug response has enormous but untapped clinical potential. Both animal and human research indicates that a wide range of drug responses, from immune suppression to motor stimulation, can be conditioned (a neutral stimulus, such as a pill or injection, associated with drug administration can in itself evoke the drug effect). In many conditioning or dose-extending models, a particular response to real medication (such as pain relief after analgesics) first becomes conditioned due to repeated exposure to the drug given in a particular vehicle. Then, the treatment shifts to some doses comprising of real medicine and some doses comprising of placebo. Because the drug response has been conditioned, it is thought that the response to an identically appearing placebo will mirror the drug response. The active drug often is only replaced by placebo for certain doses under a schedule of partial reinforcement, given the ubiquity of extinction (the conditioned response lessens when the conditioned stimulus is presented alone on repeated trials).

In 1 version of a conditioning study, children with ADHD were randomized to 1 of 3 groups.²⁸ One group (full dose) took the standard dose of medication for 2 months, a second group (reduced dose) took a standard dose during 1 month followed by a half dose during the second month, and children in the third group (reduced dose with placebo) took the standard dose plus a visually distinctive placebo during the first month, followed by a half dose plus the visually distinctive placebo during the second month. Not surprisingly, ADHD symptoms were worse among children in the reduced-dose group. However, there was no difference between those in the reduced-dose with placebo group and those in the full-dose group. It appears as though the symptom reduction associated with a 100% dose was an unconditioned response that could be mimicked with the addition of a placebo pill.

In another study, patients with psoriasis were randomly assigned to receive a full dose of active medication (0.1% triamcinolone cream) twice a day, or a full dose of active medication for 25% to 50% of the doses, with a placebo (moisturizing cream) given for the other 50% to 75% of the doses.²⁹ Relapse rates were not statistically different between groups.

“It is a mystery how a ubiquitous treatment used since antiquity was unknown, unnamed, and unidentified until recently. It is even more remarkable because this is the only treatment common to all societies and cultures.”¹

The treatment discussed above is not a specific pill, surgery, plant, or herb. Rather, the authors are referring to placebo. Indeed, the history of medical treatment is largely a chronicle of placebos. When subjected to scientific scrutiny, the overwhelming majority of treatments have turned out to be devoid of intrinsic therapeutic value; they derived their benefits from the placebo effect. Despite these benefits, the term “placebo” comes with unfortunate baggage. Latin for “I shall please,” it is the first word of the Christian vespers for the dead. In the 12th century these vespers were commonly referred to as placebos. By the 1300s, the term had become secular and pejorative, suggesting a flatterer or sycophant. When the word entered medical terminology in the late 18th century, the negative connotation stuck. A placebo was defined as a medicine given to please patients rather than to benefit them. In the modern era, the lack of pharmacologic activity became part of the definition as well.

The word placebo brings with it connotations of deception, fakery, and ineffectiveness. But one of the things about placebos that contribute mightily to the health care community’s aversion toward them is, in fact, their effectiveness. They bring relief across a wide range of medical conditions.² In doing so, placebos impugn the value of our most cherished remedies, hamper the development of new therapeutics, and threaten our livelihoods as health professionals.³

Placebos often are conceptualized as any treatment that lacks intrinsic therapeutic value, such as sugar pills. But looking at what placebo treatment actually entails, both in placebo-controlled treatment trials and in clinical settings, suggests a more comprehensive definition. Placebos encompass all the elements common to any treatment or healing situation. These include a recognized healer, evaluation, diagnosis, prognosis, plausible treatment, and most importantly, the expectation that one will recover. Along these lines, the placebo response can be thought of as the response to the common elements of the treatment or healing situation.³

Research regarding the placebo effect has mushroomed in the past 2 decades. Over this time, we have learned a good deal about both the mechanisms underlying the placebo effect and how the placebo effect can be applied to enhance the benefit of conventional treatment. Brain imaging technology has revealed that when placebo treatment alleviates pain, Parkinson’s disease, and depression, brain changes occur that are similar to those observed with active pharma­cologic treatment.^4,5 Recent studies also show that deliberate, open (nondeceptive) use of placebo can improve the symptoms of several conditions, including depression, pain, and irritable bowel syndrome.⁶ Furthermore, intermittent substitution of placebo pills for pharmacologically active treatment in a conditioning paradigm can be as effective as the “real” treatment.⁷ Also, research over the past decade has verified that certain common features of the treatment situation, particularly the quality of the doctor–patient encounter, contribute to the placebo response and have a demonstrable impact on the outcome of treatment.⁸ Clearly, the placebo effect has gone from being simply a nuisance that interferes with the evaluation of new treatments to a variable worthy of study and application in its own right. Although, for the most part, clinical practice has not kept up with these advances.

Placebos seem to have their greatest impact on the subjective symptoms of disease—pain, distress, and discouragement. It should come as no surprise, then, that placebos are particularly effective in certain psychiatric conditions. In some forms of anxiety and depressive disorders, for example, distress is the illness, and placebos reliably bring relief. Patients with panic disorder, mild to moderate depression, or generalized anxiety disorder get almost as much relief with placebo as they do with conventional treatment (about one-half improve with placebo).^9-11 But <20% of those with obsessive-compulsive disorder improve with placebo, and placebo response rates are also low in patients with schizophrenia or dementia. Mania, attention-deficit/hyperactivity disorder (ADHD), and severe depression fall somewhere in the middle.³

Harnessing the placebo response

There may be a few circumstances in psychiatric practice when it makes sense to intentionally prescribe a placebo as treatment, and we discuss those below. But far more frequently, what we know about the elements that contribute to the placebo effect can be applied to enhance the benefits of any treatment. Patients might be best served if deliberate mobilization of the placebo effect was a standard adjunct to conventional clinical care.

Various components of the treatment situation, collectively referred to as placebo, are a powerful antidote for illness, and some of these healing components exert their influence without special activity on the clinician’s part:

• Simply seeking psychiatric care can bring relief by providing some sense of control over distressing symptoms. The standard trappings of the office or clinic and customary office procedures—from the presentation of one’s insurance card to taking a history—offer reassurance and evoke the expectation that improvement or recovery is around the corner.
• The comfort provided by the psychiatrist’s presence is enhanced when patients feel that they are in the hands of a recognized healer. Psychiatrists inspire confidence when they look like a psychiatrist, or more precisely, like the patient’s idea of what a psychiatrist should look like. In our culture, that means a white coat or business attire.

A thorough evaluation is one of the common treatment elements that does the most to reduce distress and inspire confidence. The quality of an evaluation bears a strong relationship to patients’ satisfaction with the medical encounter, and can influence the amount of disability they suffer.^3,12-15

Although guidelines for conducting effective psychiatric interviews have been around for almost 100 years, psychiatrists vary considerably in the extent to which they elicit complete and accurate information, build rapport, give patients the sense that they are listened to, and provide a thorough assessment. The degree to which patients feel that the clinician is responsive to their concerns depends as much on the style of the interview as on the amount of time devoted to it. Nonverbal behavior can carry the message that the clinician is paying full attention. Something as simple as not answering the phone during an interview (this seems obvious, but a surprising and troubling number of mental health professionals take phone calls during interviews and treatment sessions) conveys an important message about the importance that the clinician places on the patient’s problems.³

The idea that the treatment situation itself provides reassurance and reduces distress, and in doing so, powers a good bit of the placebo effect, is enshrined in such concepts as the importance of good bedside manner. Many feel that the doctor’s thoughtful attention, positive regard, and optimism—so valued by patients—are justified on humanitarian grounds alone; actual evidence that this caring behavior contributes to healing isn’t required. To many, the healing properties of the treatment situation are self-evident. But as the costs of health care snowball and the demands for efficiency and cost-effectiveness rise, the time that psychiatrists can devote to patients has dwindled. Third-party payors demand evidence, beyond intuition and common sense, that diagnostic procedures and treatments have some usefulness, and rightly so.

Is there any evidence that the common components of the treatment situation provide benefit?³ More specifically, does the quality of the doctor–patient relationship and the patient’s feelings about a therapeutic encounter promote healing? Several studies suggest that the doctor–patient relationship has a demonstrable impact on symptom relief.¹⁶ In 1 study, oncologists were randomly assigned to receive a Communication Skills Training (CST) program or not. CST included a 1.5-day face-to-face workshop and 6 hours of monthly videoconferencing that focused on improving communication skills with patients.¹⁷ Lessons included building rapport, engaging in appropriate eye contact, and normalizing difficult experiences. One week after initially consulting with their physician, patients who saw an oncologist in the CST group experienced less anxiety and depression than those who saw an oncologist who did not receive CST. The benefit of CST for patient anxiety mostly persisted at a 3-month follow-up.

A recent meta-analysis pooled the results of 47 studies to examine the relationship between how much trust patients have for their doctors and health outcomes. There was a small to medium association: More trust was associated with greater improvement.¹⁸ It is possible that a good doctor–patient relationship enhances expectancies. However, it is also likely that a positive therapeutic relationship is inherently soothing and reduces distress or dysfunction independent of expectation. Regardless of the precise mechanism, these studies warrant attention. We all understand that it is important on ethical grounds to treat patients with respect and kindness. Research shows that this type of behavior also promotes recovery.

Patient expectations. The idea that expectation of improvement has a major impact on treatment outcome is firmly grounded in research on the placebo effect. Studies have shown that what people expect to experience as an outcome of treatment has a substantial impact on what they actually experience. In a classic study, a doctor told some patients with symptoms of minor illness that they would feel better soon and another group with the same symptoms that he didn’t know what ailed them.¹⁹ Two weeks later, 64% of patients in the “positive expectation” group were improved, compared with only 39% of patients in the “negative” group. In another study, adults were exposed to an allergen that caused a skin reaction.²⁰ Hand lotion (ie, a therapeutically inert substance) was then spread on the skin. Patients were led to believe that the cream would either alleviate or exacerbate the itching. The experimentally-induced wheal-and-flare was measured in both groups a few minutes after the allergen and cream were applied. The wheal-and-flare were worse for participants in the group that expected exacerbation.

Not uncommonly, expectation can have more impact on clinical outcome than a drug’s pharmacologic activity. In a double-blind placebo-controlled study, patients with depression were treated with St. John’s wort, sertraline, or placebo.²¹ They improved to the same extent with all 3 treatments. But when patients were asked to guess the treatment to which they had been assigned, those who thought they had received placebo showed little improvement, irrespective of which intervention they actually received, and those who guessed they had been given St. John’s wort or sertraline showed uniformly large improvement, irrespective of which intervention they actually received (including placebo). The researchers concluded that “Patient beliefs regarding treatment may have a stronger association with clinical outcome than the actual medication received.”

Psychiatrists who wish to use all the therapeutic tools at their disposal must attend to and manage patient expectations. One part of channeling a patient’s expectation is to thoroughly assess the patient’s beliefs regarding the efficacy of various treatments. If a patient’s uncle said that a certain drug is a miracle cure for anxiety, and the patient believes it to be true, then that expectation must be taken into consideration. Many patients prefer alternative treatments to conventional therapies. As long as there is no reason to think an alternative treatment will cause harm, a compromise might be reasonable. For example, if a patient with schizophrenia wants to treat her symptoms with herbal tea, the psychiatrist could say, “In addition to the tea, I recommend that you also take clozapine. The combination is likely to improve your symptoms.”³ More than anything else, the words a psychiatrist uses when recommending treatment shape the patient’s expectations. “You should be feeling a lot less anxious soon after you start taking this” has a different effect than “Try this. It may help.”

Prescribing ‘open-label’ placebo

There may be some limited circumstances where an actual placebo (eg, a sugar pill) might be suitable as a treatment. These include when placebo and conventional treatment provide similar results and a patient is reluctant to take conventional medicine, or when there is no effective conventional treatment. The deceptive prescription of placebo (providing placebo and calling it a drug) has a long history and was considered ethical—and recommended by medical authorities—until the latter half of the 20th century. This practice was deemed unethical in the 1980s, because it was dishonest and violated patient autonomy. Because it was widely believed that placebos given openly would be ineffective, the end of placebo treatment seemed at hand. An intriguing body of evidence, however, suggests that placebos can be effective even when patients know they are taking a placebo. Patients given an “open-label” placebo are told something along the lines of “the pill being prescribed contains no medicine, but some people improve with it, perhaps because the pill stimulates the body’s self-healing.” Open-label placebo has been evaluated for depression,²² low back pain,²³ irritable bowel syndrome,²⁴ neurosis,²⁵ allergic rhinitis,²⁶ and anxiety.²⁷ Most of these studies are small, and some were uncontrolled. Yet they consistently have shown that symptoms improve with a nondeceptive placebo, and improve to a greater extent than with no treatment.

The most recent trial is a promising example of the potential of open-label placebos. In this study, 96 patients with chronic low back pain were randomly assigned to 3 weeks of treatment as usual (TAU) or 3 weeks of TAU plus open-label placebo.²³ Patients who received open-label placebo were educated about the placebo effect and shown a film clip describing promising results of a prior open-label placebo study. They were then given placebo pills to be take once daily, and clearly told the pills contained no active medication. After 3 weeks, patients in the TAU plus placebo group reported less pain and less disability than patients who received TAU without a placebo. Some patients even requested a placebo prescription at the end of the study.

The placebo response provides a rational basis for prescribing innocuous alternative therapies with no intrinsic therapeutic value. Patients who prefer and believe in the effectiveness of alternative remedies—herbal compounds, massage, magnets, homeopathic solutions, etc.—can be recommended these treatments to mobilize a placebo response.

Using a conditioning model. Prescribing a placebo to obtain a conditioned drug response has enormous but untapped clinical potential. Both animal and human research indicates that a wide range of drug responses, from immune suppression to motor stimulation, can be conditioned (a neutral stimulus, such as a pill or injection, associated with drug administration can in itself evoke the drug effect). In many conditioning or dose-extending models, a particular response to real medication (such as pain relief after analgesics) first becomes conditioned due to repeated exposure to the drug given in a particular vehicle. Then, the treatment shifts to some doses comprising of real medicine and some doses comprising of placebo. Because the drug response has been conditioned, it is thought that the response to an identically appearing placebo will mirror the drug response. The active drug often is only replaced by placebo for certain doses under a schedule of partial reinforcement, given the ubiquity of extinction (the conditioned response lessens when the conditioned stimulus is presented alone on repeated trials).

In 1 version of a conditioning study, children with ADHD were randomized to 1 of 3 groups.²⁸ One group (full dose) took the standard dose of medication for 2 months, a second group (reduced dose) took a standard dose during 1 month followed by a half dose during the second month, and children in the third group (reduced dose with placebo) took the standard dose plus a visually distinctive placebo during the first month, followed by a half dose plus the visually distinctive placebo during the second month. Not surprisingly, ADHD symptoms were worse among children in the reduced-dose group. However, there was no difference between those in the reduced-dose with placebo group and those in the full-dose group. It appears as though the symptom reduction associated with a 100% dose was an unconditioned response that could be mimicked with the addition of a placebo pill.

In another study, patients with psoriasis were randomly assigned to receive a full dose of active medication (0.1% triamcinolone cream) twice a day, or a full dose of active medication for 25% to 50% of the doses, with a placebo (moisturizing cream) given for the other 50% to 75% of the doses.²⁹ Relapse rates were not statistically different between groups.

“It is a mystery how a ubiquitous treatment used since antiquity was unknown, unnamed, and unidentified until recently. It is even more remarkable because this is the only treatment common to all societies and cultures.”¹

The treatment discussed above is not a specific pill, surgery, plant, or herb. Rather, the authors are referring to placebo. Indeed, the history of medical treatment is largely a chronicle of placebos. When subjected to scientific scrutiny, the overwhelming majority of treatments have turned out to be devoid of intrinsic therapeutic value; they derived their benefits from the placebo effect. Despite these benefits, the term “placebo” comes with unfortunate baggage. Latin for “I shall please,” it is the first word of the Christian vespers for the dead. In the 12th century these vespers were commonly referred to as placebos. By the 1300s, the term had become secular and pejorative, suggesting a flatterer or sycophant. When the word entered medical terminology in the late 18th century, the negative connotation stuck. A placebo was defined as a medicine given to please patients rather than to benefit them. In the modern era, the lack of pharmacologic activity became part of the definition as well.

The word placebo brings with it connotations of deception, fakery, and ineffectiveness. But one of the things about placebos that contribute mightily to the health care community’s aversion toward them is, in fact, their effectiveness. They bring relief across a wide range of medical conditions.² In doing so, placebos impugn the value of our most cherished remedies, hamper the development of new therapeutics, and threaten our livelihoods as health professionals.³

Placebos often are conceptualized as any treatment that lacks intrinsic therapeutic value, such as sugar pills. But looking at what placebo treatment actually entails, both in placebo-controlled treatment trials and in clinical settings, suggests a more comprehensive definition. Placebos encompass all the elements common to any treatment or healing situation. These include a recognized healer, evaluation, diagnosis, prognosis, plausible treatment, and most importantly, the expectation that one will recover. Along these lines, the placebo response can be thought of as the response to the common elements of the treatment or healing situation.³

Research regarding the placebo effect has mushroomed in the past 2 decades. Over this time, we have learned a good deal about both the mechanisms underlying the placebo effect and how the placebo effect can be applied to enhance the benefit of conventional treatment. Brain imaging technology has revealed that when placebo treatment alleviates pain, Parkinson’s disease, and depression, brain changes occur that are similar to those observed with active pharma­cologic treatment.^4,5 Recent studies also show that deliberate, open (nondeceptive) use of placebo can improve the symptoms of several conditions, including depression, pain, and irritable bowel syndrome.⁶ Furthermore, intermittent substitution of placebo pills for pharmacologically active treatment in a conditioning paradigm can be as effective as the “real” treatment.⁷ Also, research over the past decade has verified that certain common features of the treatment situation, particularly the quality of the doctor–patient encounter, contribute to the placebo response and have a demonstrable impact on the outcome of treatment.⁸ Clearly, the placebo effect has gone from being simply a nuisance that interferes with the evaluation of new treatments to a variable worthy of study and application in its own right. Although, for the most part, clinical practice has not kept up with these advances.

Placebos seem to have their greatest impact on the subjective symptoms of disease—pain, distress, and discouragement. It should come as no surprise, then, that placebos are particularly effective in certain psychiatric conditions. In some forms of anxiety and depressive disorders, for example, distress is the illness, and placebos reliably bring relief. Patients with panic disorder, mild to moderate depression, or generalized anxiety disorder get almost as much relief with placebo as they do with conventional treatment (about one-half improve with placebo).^9-11 But <20% of those with obsessive-compulsive disorder improve with placebo, and placebo response rates are also low in patients with schizophrenia or dementia. Mania, attention-deficit/hyperactivity disorder (ADHD), and severe depression fall somewhere in the middle.³

Harnessing the placebo response

There may be a few circumstances in psychiatric practice when it makes sense to intentionally prescribe a placebo as treatment, and we discuss those below. But far more frequently, what we know about the elements that contribute to the placebo effect can be applied to enhance the benefits of any treatment. Patients might be best served if deliberate mobilization of the placebo effect was a standard adjunct to conventional clinical care.

Various components of the treatment situation, collectively referred to as placebo, are a powerful antidote for illness, and some of these healing components exert their influence without special activity on the clinician’s part:

• Simply seeking psychiatric care can bring relief by providing some sense of control over distressing symptoms. The standard trappings of the office or clinic and customary office procedures—from the presentation of one’s insurance card to taking a history—offer reassurance and evoke the expectation that improvement or recovery is around the corner.
• The comfort provided by the psychiatrist’s presence is enhanced when patients feel that they are in the hands of a recognized healer. Psychiatrists inspire confidence when they look like a psychiatrist, or more precisely, like the patient’s idea of what a psychiatrist should look like. In our culture, that means a white coat or business attire.

A thorough evaluation is one of the common treatment elements that does the most to reduce distress and inspire confidence. The quality of an evaluation bears a strong relationship to patients’ satisfaction with the medical encounter, and can influence the amount of disability they suffer.^3,12-15

Although guidelines for conducting effective psychiatric interviews have been around for almost 100 years, psychiatrists vary considerably in the extent to which they elicit complete and accurate information, build rapport, give patients the sense that they are listened to, and provide a thorough assessment. The degree to which patients feel that the clinician is responsive to their concerns depends as much on the style of the interview as on the amount of time devoted to it. Nonverbal behavior can carry the message that the clinician is paying full attention. Something as simple as not answering the phone during an interview (this seems obvious, but a surprising and troubling number of mental health professionals take phone calls during interviews and treatment sessions) conveys an important message about the importance that the clinician places on the patient’s problems.³

The idea that the treatment situation itself provides reassurance and reduces distress, and in doing so, powers a good bit of the placebo effect, is enshrined in such concepts as the importance of good bedside manner. Many feel that the doctor’s thoughtful attention, positive regard, and optimism—so valued by patients—are justified on humanitarian grounds alone; actual evidence that this caring behavior contributes to healing isn’t required. To many, the healing properties of the treatment situation are self-evident. But as the costs of health care snowball and the demands for efficiency and cost-effectiveness rise, the time that psychiatrists can devote to patients has dwindled. Third-party payors demand evidence, beyond intuition and common sense, that diagnostic procedures and treatments have some usefulness, and rightly so.

Is there any evidence that the common components of the treatment situation provide benefit?³ More specifically, does the quality of the doctor–patient relationship and the patient’s feelings about a therapeutic encounter promote healing? Several studies suggest that the doctor–patient relationship has a demonstrable impact on symptom relief.¹⁶ In 1 study, oncologists were randomly assigned to receive a Communication Skills Training (CST) program or not. CST included a 1.5-day face-to-face workshop and 6 hours of monthly videoconferencing that focused on improving communication skills with patients.¹⁷ Lessons included building rapport, engaging in appropriate eye contact, and normalizing difficult experiences. One week after initially consulting with their physician, patients who saw an oncologist in the CST group experienced less anxiety and depression than those who saw an oncologist who did not receive CST. The benefit of CST for patient anxiety mostly persisted at a 3-month follow-up.

A recent meta-analysis pooled the results of 47 studies to examine the relationship between how much trust patients have for their doctors and health outcomes. There was a small to medium association: More trust was associated with greater improvement.¹⁸ It is possible that a good doctor–patient relationship enhances expectancies. However, it is also likely that a positive therapeutic relationship is inherently soothing and reduces distress or dysfunction independent of expectation. Regardless of the precise mechanism, these studies warrant attention. We all understand that it is important on ethical grounds to treat patients with respect and kindness. Research shows that this type of behavior also promotes recovery.

Patient expectations. The idea that expectation of improvement has a major impact on treatment outcome is firmly grounded in research on the placebo effect. Studies have shown that what people expect to experience as an outcome of treatment has a substantial impact on what they actually experience. In a classic study, a doctor told some patients with symptoms of minor illness that they would feel better soon and another group with the same symptoms that he didn’t know what ailed them.¹⁹ Two weeks later, 64% of patients in the “positive expectation” group were improved, compared with only 39% of patients in the “negative” group. In another study, adults were exposed to an allergen that caused a skin reaction.²⁰ Hand lotion (ie, a therapeutically inert substance) was then spread on the skin. Patients were led to believe that the cream would either alleviate or exacerbate the itching. The experimentally-induced wheal-and-flare was measured in both groups a few minutes after the allergen and cream were applied. The wheal-and-flare were worse for participants in the group that expected exacerbation.

Not uncommonly, expectation can have more impact on clinical outcome than a drug’s pharmacologic activity. In a double-blind placebo-controlled study, patients with depression were treated with St. John’s wort, sertraline, or placebo.²¹ They improved to the same extent with all 3 treatments. But when patients were asked to guess the treatment to which they had been assigned, those who thought they had received placebo showed little improvement, irrespective of which intervention they actually received, and those who guessed they had been given St. John’s wort or sertraline showed uniformly large improvement, irrespective of which intervention they actually received (including placebo). The researchers concluded that “Patient beliefs regarding treatment may have a stronger association with clinical outcome than the actual medication received.”

Psychiatrists who wish to use all the therapeutic tools at their disposal must attend to and manage patient expectations. One part of channeling a patient’s expectation is to thoroughly assess the patient’s beliefs regarding the efficacy of various treatments. If a patient’s uncle said that a certain drug is a miracle cure for anxiety, and the patient believes it to be true, then that expectation must be taken into consideration. Many patients prefer alternative treatments to conventional therapies. As long as there is no reason to think an alternative treatment will cause harm, a compromise might be reasonable. For example, if a patient with schizophrenia wants to treat her symptoms with herbal tea, the psychiatrist could say, “In addition to the tea, I recommend that you also take clozapine. The combination is likely to improve your symptoms.”³ More than anything else, the words a psychiatrist uses when recommending treatment shape the patient’s expectations. “You should be feeling a lot less anxious soon after you start taking this” has a different effect than “Try this. It may help.”

Prescribing ‘open-label’ placebo

There may be some limited circumstances where an actual placebo (eg, a sugar pill) might be suitable as a treatment. These include when placebo and conventional treatment provide similar results and a patient is reluctant to take conventional medicine, or when there is no effective conventional treatment. The deceptive prescription of placebo (providing placebo and calling it a drug) has a long history and was considered ethical—and recommended by medical authorities—until the latter half of the 20th century. This practice was deemed unethical in the 1980s, because it was dishonest and violated patient autonomy. Because it was widely believed that placebos given openly would be ineffective, the end of placebo treatment seemed at hand. An intriguing body of evidence, however, suggests that placebos can be effective even when patients know they are taking a placebo. Patients given an “open-label” placebo are told something along the lines of “the pill being prescribed contains no medicine, but some people improve with it, perhaps because the pill stimulates the body’s self-healing.” Open-label placebo has been evaluated for depression,²² low back pain,²³ irritable bowel syndrome,²⁴ neurosis,²⁵ allergic rhinitis,²⁶ and anxiety.²⁷ Most of these studies are small, and some were uncontrolled. Yet they consistently have shown that symptoms improve with a nondeceptive placebo, and improve to a greater extent than with no treatment.

The most recent trial is a promising example of the potential of open-label placebos. In this study, 96 patients with chronic low back pain were randomly assigned to 3 weeks of treatment as usual (TAU) or 3 weeks of TAU plus open-label placebo.²³ Patients who received open-label placebo were educated about the placebo effect and shown a film clip describing promising results of a prior open-label placebo study. They were then given placebo pills to be take once daily, and clearly told the pills contained no active medication. After 3 weeks, patients in the TAU plus placebo group reported less pain and less disability than patients who received TAU without a placebo. Some patients even requested a placebo prescription at the end of the study.

The placebo response provides a rational basis for prescribing innocuous alternative therapies with no intrinsic therapeutic value. Patients who prefer and believe in the effectiveness of alternative remedies—herbal compounds, massage, magnets, homeopathic solutions, etc.—can be recommended these treatments to mobilize a placebo response.

Using a conditioning model. Prescribing a placebo to obtain a conditioned drug response has enormous but untapped clinical potential. Both animal and human research indicates that a wide range of drug responses, from immune suppression to motor stimulation, can be conditioned (a neutral stimulus, such as a pill or injection, associated with drug administration can in itself evoke the drug effect). In many conditioning or dose-extending models, a particular response to real medication (such as pain relief after analgesics) first becomes conditioned due to repeated exposure to the drug given in a particular vehicle. Then, the treatment shifts to some doses comprising of real medicine and some doses comprising of placebo. Because the drug response has been conditioned, it is thought that the response to an identically appearing placebo will mirror the drug response. The active drug often is only replaced by placebo for certain doses under a schedule of partial reinforcement, given the ubiquity of extinction (the conditioned response lessens when the conditioned stimulus is presented alone on repeated trials).

In 1 version of a conditioning study, children with ADHD were randomized to 1 of 3 groups.²⁸ One group (full dose) took the standard dose of medication for 2 months, a second group (reduced dose) took a standard dose during 1 month followed by a half dose during the second month, and children in the third group (reduced dose with placebo) took the standard dose plus a visually distinctive placebo during the first month, followed by a half dose plus the visually distinctive placebo during the second month. Not surprisingly, ADHD symptoms were worse among children in the reduced-dose group. However, there was no difference between those in the reduced-dose with placebo group and those in the full-dose group. It appears as though the symptom reduction associated with a 100% dose was an unconditioned response that could be mimicked with the addition of a placebo pill.

In another study, patients with psoriasis were randomly assigned to receive a full dose of active medication (0.1% triamcinolone cream) twice a day, or a full dose of active medication for 25% to 50% of the doses, with a placebo (moisturizing cream) given for the other 50% to 75% of the doses.²⁹ Relapse rates were not statistically different between groups.

Mr. C, age 30, with schizoaffective disorder, bipolar type, Cannabis abuse, and nicotine dependence, has been enrolled in a Program of Assertive Community Treatment (PACT) for approximately 5 years. He presents to the PACT clinic for follow-up with his psychiatrist. Mr. C reports dizziness, lightheadedness, blurred vision, and nausea worsening over the last few days, and he appears drowsy and hypoactive. He does not report any chest pain, abdominal pain, swelling, cold extremities, shortness of breath, vomiting, diarrhea, or blood loss. Mr. C admits he has eaten only once daily for several weeks because of delusional ideation that he is responsible for others suffering from anorexia nervosa.

His medical history includes gastroesophageal reflux disease. Mr. C’s medication regimen for the past year included total daily oral doses of benztropine, 2 mg, divalproex extended-release, 1,000 mg, fluphenazine, 15 mg, and gabapentin, 300 mg. He also receives IM fluphenazine decanoate, 50 mg every 2 weeks; lithium, 600 mg/d, was added to his regimen 5 months ago. Vital signs include temperature 97°F, weight 162 lb, height 69 inches, blood pressure 105/64 mm Hg, heart rate (HR) 46 beats per minute (bpm), and respirations 18 breaths per minute.

Because of Mr. C’s complaints, appearance, and low HR, the psychiatrist calls emergency medical services (EMS). Although the paramedics recommend emergency transport to the hospital, Mr. C refuses. The psychiatrist instructs Mr. C to stop taking lithium because of suspected lithium-induced bradycardia and a concern that he may be more susceptible to lithium toxicity with prolonged anorexia nervosa. When nursing staff evaluate Mr. C the next day, his vitals are HR 60 bpm, respirations 20 breaths per minute, and blood pressure 124/81 mm Hg; his dizziness, blurred vision, lightheadedness, and nausea are resolved.

Bradycardia is defined as a HR <60 bpm; however, symptoms may not occur until the HR is <50 bpm. Symptoms include fatigue, dizziness, lightheadedness, chest pain, shortness of breath, and syncope. The incidence of bradycardia during lithium treatment is unknown; it is considered a rare but serious adverse effect. A literature review reveals several case reports of bradycardia with lithium treatment,^2-4 including symptomatic bradycardia after a single dose of lithium.⁵ Other possible causes of bradycardia include anorexia nervosa, hypothermia, hypothyroidism, hypoxia, infection, stroke, acute myocardial infarction, sedative or opiate use, increased vagal tone with exercise conditioning, and other medications including fluphenazine.⁶

Mr. C’s symptoms may have been assumed to be secondary to several possible causes, including bradycardia, dehydration from poor oral intake, lithium toxicity, or an undiagnosed medical condition. The combination of nausea, dizziness, anorexia nervosa, blurred vision, and lightheadedness in a patient receiving lithium would certainly trigger a clinician’s concern for lithium toxicity, but he (she) may not be aware of the risk of bradycardia as an adverse effect of lithium. Because Mr. C refused hospital transportation by EMS, discontinuing lithium appears to have been the safest option. Laboratory studies from the day after Mr. C presented to the clinic appeared to lessen the probability that lithium toxicity, hypothyroidism, valproate toxicity, type 2 diabetes mellitus, or infection had caused Mr. C’s symptoms.

Although psychiatrists may be vigilant about monitoring for signs and symptoms of toxicity with lithium use by utilizing regular laboratory studies, they may not be as vigilant with monitoring vital signs at every patient visit (Table). This case demonstrates the importance of regular vital sign measurements to be able to detect this rare but serious adverse effect.

Mr. C, age 30, with schizoaffective disorder, bipolar type, Cannabis abuse, and nicotine dependence, has been enrolled in a Program of Assertive Community Treatment (PACT) for approximately 5 years. He presents to the PACT clinic for follow-up with his psychiatrist. Mr. C reports dizziness, lightheadedness, blurred vision, and nausea worsening over the last few days, and he appears drowsy and hypoactive. He does not report any chest pain, abdominal pain, swelling, cold extremities, shortness of breath, vomiting, diarrhea, or blood loss. Mr. C admits he has eaten only once daily for several weeks because of delusional ideation that he is responsible for others suffering from anorexia nervosa.

His medical history includes gastroesophageal reflux disease. Mr. C’s medication regimen for the past year included total daily oral doses of benztropine, 2 mg, divalproex extended-release, 1,000 mg, fluphenazine, 15 mg, and gabapentin, 300 mg. He also receives IM fluphenazine decanoate, 50 mg every 2 weeks; lithium, 600 mg/d, was added to his regimen 5 months ago. Vital signs include temperature 97°F, weight 162 lb, height 69 inches, blood pressure 105/64 mm Hg, heart rate (HR) 46 beats per minute (bpm), and respirations 18 breaths per minute.

Because of Mr. C’s complaints, appearance, and low HR, the psychiatrist calls emergency medical services (EMS). Although the paramedics recommend emergency transport to the hospital, Mr. C refuses. The psychiatrist instructs Mr. C to stop taking lithium because of suspected lithium-induced bradycardia and a concern that he may be more susceptible to lithium toxicity with prolonged anorexia nervosa. When nursing staff evaluate Mr. C the next day, his vitals are HR 60 bpm, respirations 20 breaths per minute, and blood pressure 124/81 mm Hg; his dizziness, blurred vision, lightheadedness, and nausea are resolved.

Bradycardia is defined as a HR <60 bpm; however, symptoms may not occur until the HR is <50 bpm. Symptoms include fatigue, dizziness, lightheadedness, chest pain, shortness of breath, and syncope. The incidence of bradycardia during lithium treatment is unknown; it is considered a rare but serious adverse effect. A literature review reveals several case reports of bradycardia with lithium treatment,^2-4 including symptomatic bradycardia after a single dose of lithium.⁵ Other possible causes of bradycardia include anorexia nervosa, hypothermia, hypothyroidism, hypoxia, infection, stroke, acute myocardial infarction, sedative or opiate use, increased vagal tone with exercise conditioning, and other medications including fluphenazine.⁶

Mr. C’s symptoms may have been assumed to be secondary to several possible causes, including bradycardia, dehydration from poor oral intake, lithium toxicity, or an undiagnosed medical condition. The combination of nausea, dizziness, anorexia nervosa, blurred vision, and lightheadedness in a patient receiving lithium would certainly trigger a clinician’s concern for lithium toxicity, but he (she) may not be aware of the risk of bradycardia as an adverse effect of lithium. Because Mr. C refused hospital transportation by EMS, discontinuing lithium appears to have been the safest option. Laboratory studies from the day after Mr. C presented to the clinic appeared to lessen the probability that lithium toxicity, hypothyroidism, valproate toxicity, type 2 diabetes mellitus, or infection had caused Mr. C’s symptoms.

Although psychiatrists may be vigilant about monitoring for signs and symptoms of toxicity with lithium use by utilizing regular laboratory studies, they may not be as vigilant with monitoring vital signs at every patient visit (Table). This case demonstrates the importance of regular vital sign measurements to be able to detect this rare but serious adverse effect.

Mr. C, age 30, with schizoaffective disorder, bipolar type, Cannabis abuse, and nicotine dependence, has been enrolled in a Program of Assertive Community Treatment (PACT) for approximately 5 years. He presents to the PACT clinic for follow-up with his psychiatrist. Mr. C reports dizziness, lightheadedness, blurred vision, and nausea worsening over the last few days, and he appears drowsy and hypoactive. He does not report any chest pain, abdominal pain, swelling, cold extremities, shortness of breath, vomiting, diarrhea, or blood loss. Mr. C admits he has eaten only once daily for several weeks because of delusional ideation that he is responsible for others suffering from anorexia nervosa.

His medical history includes gastroesophageal reflux disease. Mr. C’s medication regimen for the past year included total daily oral doses of benztropine, 2 mg, divalproex extended-release, 1,000 mg, fluphenazine, 15 mg, and gabapentin, 300 mg. He also receives IM fluphenazine decanoate, 50 mg every 2 weeks; lithium, 600 mg/d, was added to his regimen 5 months ago. Vital signs include temperature 97°F, weight 162 lb, height 69 inches, blood pressure 105/64 mm Hg, heart rate (HR) 46 beats per minute (bpm), and respirations 18 breaths per minute.

Because of Mr. C’s complaints, appearance, and low HR, the psychiatrist calls emergency medical services (EMS). Although the paramedics recommend emergency transport to the hospital, Mr. C refuses. The psychiatrist instructs Mr. C to stop taking lithium because of suspected lithium-induced bradycardia and a concern that he may be more susceptible to lithium toxicity with prolonged anorexia nervosa. When nursing staff evaluate Mr. C the next day, his vitals are HR 60 bpm, respirations 20 breaths per minute, and blood pressure 124/81 mm Hg; his dizziness, blurred vision, lightheadedness, and nausea are resolved.

Bradycardia is defined as a HR <60 bpm; however, symptoms may not occur until the HR is <50 bpm. Symptoms include fatigue, dizziness, lightheadedness, chest pain, shortness of breath, and syncope. The incidence of bradycardia during lithium treatment is unknown; it is considered a rare but serious adverse effect. A literature review reveals several case reports of bradycardia with lithium treatment,^2-4 including symptomatic bradycardia after a single dose of lithium.⁵ Other possible causes of bradycardia include anorexia nervosa, hypothermia, hypothyroidism, hypoxia, infection, stroke, acute myocardial infarction, sedative or opiate use, increased vagal tone with exercise conditioning, and other medications including fluphenazine.⁶

Mr. C’s symptoms may have been assumed to be secondary to several possible causes, including bradycardia, dehydration from poor oral intake, lithium toxicity, or an undiagnosed medical condition. The combination of nausea, dizziness, anorexia nervosa, blurred vision, and lightheadedness in a patient receiving lithium would certainly trigger a clinician’s concern for lithium toxicity, but he (she) may not be aware of the risk of bradycardia as an adverse effect of lithium. Because Mr. C refused hospital transportation by EMS, discontinuing lithium appears to have been the safest option. Laboratory studies from the day after Mr. C presented to the clinic appeared to lessen the probability that lithium toxicity, hypothyroidism, valproate toxicity, type 2 diabetes mellitus, or infection had caused Mr. C’s symptoms.

Although psychiatrists may be vigilant about monitoring for signs and symptoms of toxicity with lithium use by utilizing regular laboratory studies, they may not be as vigilant with monitoring vital signs at every patient visit (Table). This case demonstrates the importance of regular vital sign measurements to be able to detect this rare but serious adverse effect.